diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,16120 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 10.0,
+  "eval_steps": 100,
+  "global_step": 24030,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.008322929671244279,
+      "grad_norm": 0.0830078125,
+      "learning_rate": 2.08073241781107e-08,
+      "loss": 0.2994,
+      "step": 20
+    },
+    {
+      "epoch": 0.016645859342488557,
+      "grad_norm": 0.0849609375,
+      "learning_rate": 4.16146483562214e-08,
+      "loss": 0.2962,
+      "step": 40
+    },
+    {
+      "epoch": 0.024968789013732832,
+      "grad_norm": 0.0947265625,
+      "learning_rate": 6.242197253433209e-08,
+      "loss": 0.2974,
+      "step": 60
+    },
+    {
+      "epoch": 0.033291718684977115,
+      "grad_norm": 0.08544921875,
+      "learning_rate": 8.32292967124428e-08,
+      "loss": 0.3034,
+      "step": 80
+    },
+    {
+      "epoch": 0.04161464835622139,
+      "grad_norm": 0.07861328125,
+      "learning_rate": 1.0403662089055348e-07,
+      "loss": 0.2985,
+      "step": 100
+    },
+    {
+      "epoch": 0.04161464835622139,
+      "eval_main_loss": 0.3075891137123108,
+      "eval_main_runtime": 6.3529,
+      "eval_main_samples_per_second": 29.908,
+      "eval_main_steps_per_second": 3.778,
+      "step": 100
+    },
+    {
+      "epoch": 0.04161464835622139,
+      "eval_anatomy_loss": 2.9698662757873535,
+      "eval_anatomy_runtime": 0.2668,
+      "eval_anatomy_samples_per_second": 7.496,
+      "eval_anatomy_steps_per_second": 3.748,
+      "step": 100
+    },
+    {
+      "epoch": 0.04161464835622139,
+      "eval_college_mathematics_loss": 2.178929567337036,
+      "eval_college_mathematics_runtime": 0.2663,
+      "eval_college_mathematics_samples_per_second": 7.511,
+      "eval_college_mathematics_steps_per_second": 3.755,
+      "step": 100
+    },
+    {
+      "epoch": 0.04161464835622139,
+      "eval_international_law_loss": 3.19402813911438,
+      "eval_international_law_runtime": 0.2669,
+      "eval_international_law_samples_per_second": 7.493,
+      "eval_international_law_steps_per_second": 3.746,
+      "step": 100
+    },
+    {
+      "epoch": 0.049937578027465665,
+      "grad_norm": 0.076171875,
+      "learning_rate": 1.2484394506866418e-07,
+      "loss": 0.2971,
+      "step": 120
+    },
+    {
+      "epoch": 0.05826050769870995,
+      "grad_norm": 0.08349609375,
+      "learning_rate": 1.4565126924677488e-07,
+      "loss": 0.2983,
+      "step": 140
+    },
+    {
+      "epoch": 0.06658343736995423,
+      "grad_norm": 0.083984375,
+      "learning_rate": 1.664585934248856e-07,
+      "loss": 0.2998,
+      "step": 160
+    },
+    {
+      "epoch": 0.0749063670411985,
+      "grad_norm": 0.0791015625,
+      "learning_rate": 1.8726591760299626e-07,
+      "loss": 0.2974,
+      "step": 180
+    },
+    {
+      "epoch": 0.08322929671244278,
+      "grad_norm": 0.07568359375,
+      "learning_rate": 2.0807324178110696e-07,
+      "loss": 0.2981,
+      "step": 200
+    },
+    {
+      "epoch": 0.08322929671244278,
+      "eval_main_loss": 0.307395339012146,
+      "eval_main_runtime": 6.3616,
+      "eval_main_samples_per_second": 29.866,
+      "eval_main_steps_per_second": 3.773,
+      "step": 200
+    },
+    {
+      "epoch": 0.08322929671244278,
+      "eval_anatomy_loss": 2.9686527252197266,
+      "eval_anatomy_runtime": 0.2674,
+      "eval_anatomy_samples_per_second": 7.48,
+      "eval_anatomy_steps_per_second": 3.74,
+      "step": 200
+    },
+    {
+      "epoch": 0.08322929671244278,
+      "eval_college_mathematics_loss": 2.1776983737945557,
+      "eval_college_mathematics_runtime": 0.2667,
+      "eval_college_mathematics_samples_per_second": 7.499,
+      "eval_college_mathematics_steps_per_second": 3.75,
+      "step": 200
+    },
+    {
+      "epoch": 0.08322929671244278,
+      "eval_international_law_loss": 3.1930840015411377,
+      "eval_international_law_runtime": 0.268,
+      "eval_international_law_samples_per_second": 7.463,
+      "eval_international_law_steps_per_second": 3.732,
+      "step": 200
+    },
+    {
+      "epoch": 0.09155222638368705,
+      "grad_norm": 0.08251953125,
+      "learning_rate": 2.2888056595921765e-07,
+      "loss": 0.3004,
+      "step": 220
+    },
+    {
+      "epoch": 0.09987515605493133,
+      "grad_norm": 0.08837890625,
+      "learning_rate": 2.4968789013732837e-07,
+      "loss": 0.3023,
+      "step": 240
+    },
+    {
+      "epoch": 0.10819808572617562,
+      "grad_norm": 0.0888671875,
+      "learning_rate": 2.704952143154391e-07,
+      "loss": 0.3018,
+      "step": 260
+    },
+    {
+      "epoch": 0.1165210153974199,
+      "grad_norm": 0.09130859375,
+      "learning_rate": 2.9130253849354976e-07,
+      "loss": 0.3018,
+      "step": 280
+    },
+    {
+      "epoch": 0.12484394506866417,
+      "grad_norm": 0.08642578125,
+      "learning_rate": 3.121098626716604e-07,
+      "loss": 0.2971,
+      "step": 300
+    },
+    {
+      "epoch": 0.12484394506866417,
+      "eval_main_loss": 0.3072234094142914,
+      "eval_main_runtime": 6.363,
+      "eval_main_samples_per_second": 29.86,
+      "eval_main_steps_per_second": 3.772,
+      "step": 300
+    },
+    {
+      "epoch": 0.12484394506866417,
+      "eval_anatomy_loss": 2.9682118892669678,
+      "eval_anatomy_runtime": 0.2675,
+      "eval_anatomy_samples_per_second": 7.476,
+      "eval_anatomy_steps_per_second": 3.738,
+      "step": 300
+    },
+    {
+      "epoch": 0.12484394506866417,
+      "eval_college_mathematics_loss": 2.1782639026641846,
+      "eval_college_mathematics_runtime": 0.267,
+      "eval_college_mathematics_samples_per_second": 7.492,
+      "eval_college_mathematics_steps_per_second": 3.746,
+      "step": 300
+    },
+    {
+      "epoch": 0.12484394506866417,
+      "eval_international_law_loss": 3.1912684440612793,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.505,
+      "eval_international_law_steps_per_second": 3.752,
+      "step": 300
+    },
+    {
+      "epoch": 0.13316687473990846,
+      "grad_norm": 0.087890625,
+      "learning_rate": 3.329171868497712e-07,
+      "loss": 0.3034,
+      "step": 320
+    },
+    {
+      "epoch": 0.14148980441115272,
+      "grad_norm": 0.08837890625,
+      "learning_rate": 3.537245110278818e-07,
+      "loss": 0.299,
+      "step": 340
+    },
+    {
+      "epoch": 0.149812734082397,
+      "grad_norm": 0.08251953125,
+      "learning_rate": 3.7453183520599253e-07,
+      "loss": 0.2994,
+      "step": 360
+    },
+    {
+      "epoch": 0.15813566375364127,
+      "grad_norm": 0.09326171875,
+      "learning_rate": 3.953391593841032e-07,
+      "loss": 0.3001,
+      "step": 380
+    },
+    {
+      "epoch": 0.16645859342488556,
+      "grad_norm": 0.1025390625,
+      "learning_rate": 4.161464835622139e-07,
+      "loss": 0.2971,
+      "step": 400
+    },
+    {
+      "epoch": 0.16645859342488556,
+      "eval_main_loss": 0.3068091869354248,
+      "eval_main_runtime": 6.3609,
+      "eval_main_samples_per_second": 29.87,
+      "eval_main_steps_per_second": 3.773,
+      "step": 400
+    },
+    {
+      "epoch": 0.16645859342488556,
+      "eval_anatomy_loss": 2.969444990158081,
+      "eval_anatomy_runtime": 0.2677,
+      "eval_anatomy_samples_per_second": 7.472,
+      "eval_anatomy_steps_per_second": 3.736,
+      "step": 400
+    },
+    {
+      "epoch": 0.16645859342488556,
+      "eval_college_mathematics_loss": 2.1761982440948486,
+      "eval_college_mathematics_runtime": 0.2684,
+      "eval_college_mathematics_samples_per_second": 7.45,
+      "eval_college_mathematics_steps_per_second": 3.725,
+      "step": 400
+    },
+    {
+      "epoch": 0.16645859342488556,
+      "eval_international_law_loss": 3.1895108222961426,
+      "eval_international_law_runtime": 0.2681,
+      "eval_international_law_samples_per_second": 7.461,
+      "eval_international_law_steps_per_second": 3.731,
+      "step": 400
+    },
+    {
+      "epoch": 0.17478152309612985,
+      "grad_norm": 0.11279296875,
+      "learning_rate": 4.3695380774032463e-07,
+      "loss": 0.2971,
+      "step": 420
+    },
+    {
+      "epoch": 0.1831044527673741,
+      "grad_norm": 0.10400390625,
+      "learning_rate": 4.577611319184353e-07,
+      "loss": 0.3002,
+      "step": 440
+    },
+    {
+      "epoch": 0.1914273824386184,
+      "grad_norm": 0.10498046875,
+      "learning_rate": 4.785684560965461e-07,
+      "loss": 0.3001,
+      "step": 460
+    },
+    {
+      "epoch": 0.19975031210986266,
+      "grad_norm": 0.1171875,
+      "learning_rate": 4.993757802746567e-07,
+      "loss": 0.2933,
+      "step": 480
+    },
+    {
+      "epoch": 0.20807324178110695,
+      "grad_norm": 0.11767578125,
+      "learning_rate": 5.201831044527674e-07,
+      "loss": 0.2953,
+      "step": 500
+    },
+    {
+      "epoch": 0.20807324178110695,
+      "eval_main_loss": 0.3060111999511719,
+      "eval_main_runtime": 6.3306,
+      "eval_main_samples_per_second": 30.013,
+      "eval_main_steps_per_second": 3.791,
+      "step": 500
+    },
+    {
+      "epoch": 0.20807324178110695,
+      "eval_anatomy_loss": 2.966276168823242,
+      "eval_anatomy_runtime": 0.2661,
+      "eval_anatomy_samples_per_second": 7.515,
+      "eval_anatomy_steps_per_second": 3.758,
+      "step": 500
+    },
+    {
+      "epoch": 0.20807324178110695,
+      "eval_college_mathematics_loss": 2.173703193664551,
+      "eval_college_mathematics_runtime": 0.2668,
+      "eval_college_mathematics_samples_per_second": 7.498,
+      "eval_college_mathematics_steps_per_second": 3.749,
+      "step": 500
+    },
+    {
+      "epoch": 0.20807324178110695,
+      "eval_international_law_loss": 3.1900217533111572,
+      "eval_international_law_runtime": 0.2668,
+      "eval_international_law_samples_per_second": 7.496,
+      "eval_international_law_steps_per_second": 3.748,
+      "step": 500
+    },
+    {
+      "epoch": 0.21639617145235124,
+      "grad_norm": 0.11962890625,
+      "learning_rate": 5.409904286308782e-07,
+      "loss": 0.3011,
+      "step": 520
+    },
+    {
+      "epoch": 0.2247191011235955,
+      "grad_norm": 0.12890625,
+      "learning_rate": 5.617977528089888e-07,
+      "loss": 0.2986,
+      "step": 540
+    },
+    {
+      "epoch": 0.2330420307948398,
+      "grad_norm": 0.142578125,
+      "learning_rate": 5.826050769870995e-07,
+      "loss": 0.3017,
+      "step": 560
+    },
+    {
+      "epoch": 0.24136496046608405,
+      "grad_norm": 0.1455078125,
+      "learning_rate": 6.034124011652102e-07,
+      "loss": 0.2957,
+      "step": 580
+    },
+    {
+      "epoch": 0.24968789013732834,
+      "grad_norm": 0.166015625,
+      "learning_rate": 6.242197253433208e-07,
+      "loss": 0.2969,
+      "step": 600
+    },
+    {
+      "epoch": 0.24968789013732834,
+      "eval_main_loss": 0.3044416904449463,
+      "eval_main_runtime": 6.333,
+      "eval_main_samples_per_second": 30.002,
+      "eval_main_steps_per_second": 3.79,
+      "step": 600
+    },
+    {
+      "epoch": 0.24968789013732834,
+      "eval_anatomy_loss": 2.9617347717285156,
+      "eval_anatomy_runtime": 0.2665,
+      "eval_anatomy_samples_per_second": 7.505,
+      "eval_anatomy_steps_per_second": 3.753,
+      "step": 600
+    },
+    {
+      "epoch": 0.24968789013732834,
+      "eval_college_mathematics_loss": 2.169576644897461,
+      "eval_college_mathematics_runtime": 0.2655,
+      "eval_college_mathematics_samples_per_second": 7.534,
+      "eval_college_mathematics_steps_per_second": 3.767,
+      "step": 600
+    },
+    {
+      "epoch": 0.24968789013732834,
+      "eval_international_law_loss": 3.184370756149292,
+      "eval_international_law_runtime": 0.2656,
+      "eval_international_law_samples_per_second": 7.53,
+      "eval_international_law_steps_per_second": 3.765,
+      "step": 600
+    },
+    {
+      "epoch": 0.2580108198085726,
+      "grad_norm": 0.16796875,
+      "learning_rate": 6.450270495214315e-07,
+      "loss": 0.2953,
+      "step": 620
+    },
+    {
+      "epoch": 0.2663337494798169,
+      "grad_norm": 0.1650390625,
+      "learning_rate": 6.658343736995424e-07,
+      "loss": 0.2994,
+      "step": 640
+    },
+    {
+      "epoch": 0.2746566791510612,
+      "grad_norm": 0.1796875,
+      "learning_rate": 6.86641697877653e-07,
+      "loss": 0.2941,
+      "step": 660
+    },
+    {
+      "epoch": 0.28297960882230544,
+      "grad_norm": 0.1611328125,
+      "learning_rate": 7.074490220557636e-07,
+      "loss": 0.2982,
+      "step": 680
+    },
+    {
+      "epoch": 0.29130253849354976,
+      "grad_norm": 0.1845703125,
+      "learning_rate": 7.282563462338745e-07,
+      "loss": 0.2956,
+      "step": 700
+    },
+    {
+      "epoch": 0.29130253849354976,
+      "eval_main_loss": 0.3035320043563843,
+      "eval_main_runtime": 6.3633,
+      "eval_main_samples_per_second": 29.859,
+      "eval_main_steps_per_second": 3.772,
+      "step": 700
+    },
+    {
+      "epoch": 0.29130253849354976,
+      "eval_anatomy_loss": 2.9591825008392334,
+      "eval_anatomy_runtime": 0.2671,
+      "eval_anatomy_samples_per_second": 7.487,
+      "eval_anatomy_steps_per_second": 3.744,
+      "step": 700
+    },
+    {
+      "epoch": 0.29130253849354976,
+      "eval_college_mathematics_loss": 2.1641225814819336,
+      "eval_college_mathematics_runtime": 0.267,
+      "eval_college_mathematics_samples_per_second": 7.491,
+      "eval_college_mathematics_steps_per_second": 3.746,
+      "step": 700
+    },
+    {
+      "epoch": 0.29130253849354976,
+      "eval_international_law_loss": 3.183608293533325,
+      "eval_international_law_runtime": 0.2677,
+      "eval_international_law_samples_per_second": 7.47,
+      "eval_international_law_steps_per_second": 3.735,
+      "step": 700
+    },
+    {
+      "epoch": 0.299625468164794,
+      "grad_norm": 0.1806640625,
+      "learning_rate": 7.490636704119851e-07,
+      "loss": 0.2968,
+      "step": 720
+    },
+    {
+      "epoch": 0.3079483978360383,
+      "grad_norm": 0.1748046875,
+      "learning_rate": 7.698709945900957e-07,
+      "loss": 0.2924,
+      "step": 740
+    },
+    {
+      "epoch": 0.31627132750728254,
+      "grad_norm": 0.1787109375,
+      "learning_rate": 7.906783187682064e-07,
+      "loss": 0.2948,
+      "step": 760
+    },
+    {
+      "epoch": 0.32459425717852686,
+      "grad_norm": 0.173828125,
+      "learning_rate": 8.114856429463172e-07,
+      "loss": 0.2987,
+      "step": 780
+    },
+    {
+      "epoch": 0.3329171868497711,
+      "grad_norm": 0.16015625,
+      "learning_rate": 8.322929671244278e-07,
+      "loss": 0.3004,
+      "step": 800
+    },
+    {
+      "epoch": 0.3329171868497711,
+      "eval_main_loss": 0.3030804991722107,
+      "eval_main_runtime": 6.3678,
+      "eval_main_samples_per_second": 29.838,
+      "eval_main_steps_per_second": 3.769,
+      "step": 800
+    },
+    {
+      "epoch": 0.3329171868497711,
+      "eval_anatomy_loss": 2.958465337753296,
+      "eval_anatomy_runtime": 0.2675,
+      "eval_anatomy_samples_per_second": 7.476,
+      "eval_anatomy_steps_per_second": 3.738,
+      "step": 800
+    },
+    {
+      "epoch": 0.3329171868497711,
+      "eval_college_mathematics_loss": 2.1627776622772217,
+      "eval_college_mathematics_runtime": 0.2676,
+      "eval_college_mathematics_samples_per_second": 7.474,
+      "eval_college_mathematics_steps_per_second": 3.737,
+      "step": 800
+    },
+    {
+      "epoch": 0.3329171868497711,
+      "eval_international_law_loss": 3.1812000274658203,
+      "eval_international_law_runtime": 0.2664,
+      "eval_international_law_samples_per_second": 7.507,
+      "eval_international_law_steps_per_second": 3.754,
+      "step": 800
+    },
+    {
+      "epoch": 0.3412401165210154,
+      "grad_norm": 0.1650390625,
+      "learning_rate": 8.531002913025385e-07,
+      "loss": 0.2941,
+      "step": 820
+    },
+    {
+      "epoch": 0.3495630461922597,
+      "grad_norm": 0.1806640625,
+      "learning_rate": 8.739076154806493e-07,
+      "loss": 0.2961,
+      "step": 840
+    },
+    {
+      "epoch": 0.35788597586350396,
+      "grad_norm": 0.189453125,
+      "learning_rate": 8.947149396587599e-07,
+      "loss": 0.2957,
+      "step": 860
+    },
+    {
+      "epoch": 0.3662089055347482,
+      "grad_norm": 0.1708984375,
+      "learning_rate": 9.155222638368706e-07,
+      "loss": 0.2941,
+      "step": 880
+    },
+    {
+      "epoch": 0.37453183520599254,
+      "grad_norm": 0.203125,
+      "learning_rate": 9.363295880149814e-07,
+      "loss": 0.2956,
+      "step": 900
+    },
+    {
+      "epoch": 0.37453183520599254,
+      "eval_main_loss": 0.30232974886894226,
+      "eval_main_runtime": 6.3643,
+      "eval_main_samples_per_second": 29.854,
+      "eval_main_steps_per_second": 3.771,
+      "step": 900
+    },
+    {
+      "epoch": 0.37453183520599254,
+      "eval_anatomy_loss": 2.9550912380218506,
+      "eval_anatomy_runtime": 0.2674,
+      "eval_anatomy_samples_per_second": 7.478,
+      "eval_anatomy_steps_per_second": 3.739,
+      "step": 900
+    },
+    {
+      "epoch": 0.37453183520599254,
+      "eval_college_mathematics_loss": 2.162787914276123,
+      "eval_college_mathematics_runtime": 0.266,
+      "eval_college_mathematics_samples_per_second": 7.518,
+      "eval_college_mathematics_steps_per_second": 3.759,
+      "step": 900
+    },
+    {
+      "epoch": 0.37453183520599254,
+      "eval_international_law_loss": 3.1821956634521484,
+      "eval_international_law_runtime": 0.2683,
+      "eval_international_law_samples_per_second": 7.456,
+      "eval_international_law_steps_per_second": 3.728,
+      "step": 900
+    },
+    {
+      "epoch": 0.3828547648772368,
+      "grad_norm": 0.21875,
+      "learning_rate": 9.571369121930921e-07,
+      "loss": 0.2989,
+      "step": 920
+    },
+    {
+      "epoch": 0.39117769454848106,
+      "grad_norm": 0.23046875,
+      "learning_rate": 9.779442363712028e-07,
+      "loss": 0.2965,
+      "step": 940
+    },
+    {
+      "epoch": 0.3995006242197253,
+      "grad_norm": 0.2001953125,
+      "learning_rate": 9.987515605493135e-07,
+      "loss": 0.2969,
+      "step": 960
+    },
+    {
+      "epoch": 0.40782355389096964,
+      "grad_norm": 0.2275390625,
+      "learning_rate": 1.0195588847274241e-06,
+      "loss": 0.2913,
+      "step": 980
+    },
+    {
+      "epoch": 0.4161464835622139,
+      "grad_norm": 0.275390625,
+      "learning_rate": 1.0403662089055348e-06,
+      "loss": 0.2945,
+      "step": 1000
+    },
+    {
+      "epoch": 0.4161464835622139,
+      "eval_main_loss": 0.3009730279445648,
+      "eval_main_runtime": 6.3379,
+      "eval_main_samples_per_second": 29.978,
+      "eval_main_steps_per_second": 3.787,
+      "step": 1000
+    },
+    {
+      "epoch": 0.4161464835622139,
+      "eval_anatomy_loss": 2.9502501487731934,
+      "eval_anatomy_runtime": 0.2662,
+      "eval_anatomy_samples_per_second": 7.512,
+      "eval_anatomy_steps_per_second": 3.756,
+      "step": 1000
+    },
+    {
+      "epoch": 0.4161464835622139,
+      "eval_college_mathematics_loss": 2.1539924144744873,
+      "eval_college_mathematics_runtime": 0.2669,
+      "eval_college_mathematics_samples_per_second": 7.493,
+      "eval_college_mathematics_steps_per_second": 3.746,
+      "step": 1000
+    },
+    {
+      "epoch": 0.4161464835622139,
+      "eval_international_law_loss": 3.1735572814941406,
+      "eval_international_law_runtime": 0.268,
+      "eval_international_law_samples_per_second": 7.464,
+      "eval_international_law_steps_per_second": 3.732,
+      "step": 1000
+    },
+    {
+      "epoch": 0.42446941323345816,
+      "grad_norm": 0.2294921875,
+      "learning_rate": 1.0611735330836455e-06,
+      "loss": 0.2908,
+      "step": 1020
+    },
+    {
+      "epoch": 0.4327923429047025,
+      "grad_norm": 0.26953125,
+      "learning_rate": 1.0819808572617564e-06,
+      "loss": 0.2923,
+      "step": 1040
+    },
+    {
+      "epoch": 0.44111527257594674,
+      "grad_norm": 0.25390625,
+      "learning_rate": 1.102788181439867e-06,
+      "loss": 0.291,
+      "step": 1060
+    },
+    {
+      "epoch": 0.449438202247191,
+      "grad_norm": 0.255859375,
+      "learning_rate": 1.1235955056179777e-06,
+      "loss": 0.2914,
+      "step": 1080
+    },
+    {
+      "epoch": 0.4577611319184353,
+      "grad_norm": 0.275390625,
+      "learning_rate": 1.1444028297960884e-06,
+      "loss": 0.2921,
+      "step": 1100
+    },
+    {
+      "epoch": 0.4577611319184353,
+      "eval_main_loss": 0.29984337091445923,
+      "eval_main_runtime": 6.3525,
+      "eval_main_samples_per_second": 29.909,
+      "eval_main_steps_per_second": 3.778,
+      "step": 1100
+    },
+    {
+      "epoch": 0.4577611319184353,
+      "eval_anatomy_loss": 2.947199821472168,
+      "eval_anatomy_runtime": 0.27,
+      "eval_anatomy_samples_per_second": 7.407,
+      "eval_anatomy_steps_per_second": 3.703,
+      "step": 1100
+    },
+    {
+      "epoch": 0.4577611319184353,
+      "eval_college_mathematics_loss": 2.1513562202453613,
+      "eval_college_mathematics_runtime": 0.2683,
+      "eval_college_mathematics_samples_per_second": 7.454,
+      "eval_college_mathematics_steps_per_second": 3.727,
+      "step": 1100
+    },
+    {
+      "epoch": 0.4577611319184353,
+      "eval_international_law_loss": 3.1728124618530273,
+      "eval_international_law_runtime": 0.2664,
+      "eval_international_law_samples_per_second": 7.507,
+      "eval_international_law_steps_per_second": 3.754,
+      "step": 1100
+    },
+    {
+      "epoch": 0.4660840615896796,
+      "grad_norm": 0.271484375,
+      "learning_rate": 1.165210153974199e-06,
+      "loss": 0.2917,
+      "step": 1120
+    },
+    {
+      "epoch": 0.47440699126092384,
+      "grad_norm": 0.30078125,
+      "learning_rate": 1.1860174781523097e-06,
+      "loss": 0.2874,
+      "step": 1140
+    },
+    {
+      "epoch": 0.4827299209321681,
+      "grad_norm": 0.298828125,
+      "learning_rate": 1.2068248023304204e-06,
+      "loss": 0.2926,
+      "step": 1160
+    },
+    {
+      "epoch": 0.4910528506034124,
+      "grad_norm": 0.294921875,
+      "learning_rate": 1.2276321265085312e-06,
+      "loss": 0.2913,
+      "step": 1180
+    },
+    {
+      "epoch": 0.4993757802746567,
+      "grad_norm": 0.2890625,
+      "learning_rate": 1.2484394506866417e-06,
+      "loss": 0.2926,
+      "step": 1200
+    },
+    {
+      "epoch": 0.4993757802746567,
+      "eval_main_loss": 0.29899466037750244,
+      "eval_main_runtime": 6.3585,
+      "eval_main_samples_per_second": 29.881,
+      "eval_main_steps_per_second": 3.774,
+      "step": 1200
+    },
+    {
+      "epoch": 0.4993757802746567,
+      "eval_anatomy_loss": 2.9453182220458984,
+      "eval_anatomy_runtime": 0.2675,
+      "eval_anatomy_samples_per_second": 7.477,
+      "eval_anatomy_steps_per_second": 3.738,
+      "step": 1200
+    },
+    {
+      "epoch": 0.4993757802746567,
+      "eval_college_mathematics_loss": 2.1483311653137207,
+      "eval_college_mathematics_runtime": 0.2675,
+      "eval_college_mathematics_samples_per_second": 7.477,
+      "eval_college_mathematics_steps_per_second": 3.739,
+      "step": 1200
+    },
+    {
+      "epoch": 0.4993757802746567,
+      "eval_international_law_loss": 3.1692144870758057,
+      "eval_international_law_runtime": 0.2669,
+      "eval_international_law_samples_per_second": 7.493,
+      "eval_international_law_steps_per_second": 3.746,
+      "step": 1200
+    },
+    {
+      "epoch": 0.5076987099459009,
+      "grad_norm": 0.294921875,
+      "learning_rate": 1.2692467748647524e-06,
+      "loss": 0.2921,
+      "step": 1220
+    },
+    {
+      "epoch": 0.5160216396171452,
+      "grad_norm": 0.29296875,
+      "learning_rate": 1.290054099042863e-06,
+      "loss": 0.2888,
+      "step": 1240
+    },
+    {
+      "epoch": 0.5243445692883895,
+      "grad_norm": 0.2734375,
+      "learning_rate": 1.3108614232209737e-06,
+      "loss": 0.2914,
+      "step": 1260
+    },
+    {
+      "epoch": 0.5326674989596338,
+      "grad_norm": 0.306640625,
+      "learning_rate": 1.3316687473990848e-06,
+      "loss": 0.2912,
+      "step": 1280
+    },
+    {
+      "epoch": 0.5409904286308781,
+      "grad_norm": 0.2890625,
+      "learning_rate": 1.3524760715771954e-06,
+      "loss": 0.2909,
+      "step": 1300
+    },
+    {
+      "epoch": 0.5409904286308781,
+      "eval_main_loss": 0.29816609621047974,
+      "eval_main_runtime": 6.3672,
+      "eval_main_samples_per_second": 29.84,
+      "eval_main_steps_per_second": 3.769,
+      "step": 1300
+    },
+    {
+      "epoch": 0.5409904286308781,
+      "eval_anatomy_loss": 2.942121744155884,
+      "eval_anatomy_runtime": 0.2697,
+      "eval_anatomy_samples_per_second": 7.416,
+      "eval_anatomy_steps_per_second": 3.708,
+      "step": 1300
+    },
+    {
+      "epoch": 0.5409904286308781,
+      "eval_college_mathematics_loss": 2.1458935737609863,
+      "eval_college_mathematics_runtime": 0.2675,
+      "eval_college_mathematics_samples_per_second": 7.476,
+      "eval_college_mathematics_steps_per_second": 3.738,
+      "step": 1300
+    },
+    {
+      "epoch": 0.5409904286308781,
+      "eval_international_law_loss": 3.1678075790405273,
+      "eval_international_law_runtime": 0.2673,
+      "eval_international_law_samples_per_second": 7.482,
+      "eval_international_law_steps_per_second": 3.741,
+      "step": 1300
+    },
+    {
+      "epoch": 0.5493133583021224,
+      "grad_norm": 0.291015625,
+      "learning_rate": 1.373283395755306e-06,
+      "loss": 0.2928,
+      "step": 1320
+    },
+    {
+      "epoch": 0.5576362879733666,
+      "grad_norm": 0.263671875,
+      "learning_rate": 1.3940907199334166e-06,
+      "loss": 0.2915,
+      "step": 1340
+    },
+    {
+      "epoch": 0.5659592176446109,
+      "grad_norm": 0.26953125,
+      "learning_rate": 1.4148980441115272e-06,
+      "loss": 0.2893,
+      "step": 1360
+    },
+    {
+      "epoch": 0.5742821473158551,
+      "grad_norm": 0.283203125,
+      "learning_rate": 1.435705368289638e-06,
+      "loss": 0.2907,
+      "step": 1380
+    },
+    {
+      "epoch": 0.5826050769870995,
+      "grad_norm": 0.28125,
+      "learning_rate": 1.456512692467749e-06,
+      "loss": 0.2938,
+      "step": 1400
+    },
+    {
+      "epoch": 0.5826050769870995,
+      "eval_main_loss": 0.2975090444087982,
+      "eval_main_runtime": 6.3608,
+      "eval_main_samples_per_second": 29.87,
+      "eval_main_steps_per_second": 3.773,
+      "step": 1400
+    },
+    {
+      "epoch": 0.5826050769870995,
+      "eval_anatomy_loss": 2.940919876098633,
+      "eval_anatomy_runtime": 0.2668,
+      "eval_anatomy_samples_per_second": 7.495,
+      "eval_anatomy_steps_per_second": 3.747,
+      "step": 1400
+    },
+    {
+      "epoch": 0.5826050769870995,
+      "eval_college_mathematics_loss": 2.1454479694366455,
+      "eval_college_mathematics_runtime": 0.2668,
+      "eval_college_mathematics_samples_per_second": 7.495,
+      "eval_college_mathematics_steps_per_second": 3.748,
+      "step": 1400
+    },
+    {
+      "epoch": 0.5826050769870995,
+      "eval_international_law_loss": 3.1660165786743164,
+      "eval_international_law_runtime": 0.2674,
+      "eval_international_law_samples_per_second": 7.48,
+      "eval_international_law_steps_per_second": 3.74,
+      "step": 1400
+    },
+    {
+      "epoch": 0.5909280066583438,
+      "grad_norm": 0.310546875,
+      "learning_rate": 1.4773200166458597e-06,
+      "loss": 0.2922,
+      "step": 1420
+    },
+    {
+      "epoch": 0.599250936329588,
+      "grad_norm": 0.29296875,
+      "learning_rate": 1.4981273408239701e-06,
+      "loss": 0.2957,
+      "step": 1440
+    },
+    {
+      "epoch": 0.6075738660008323,
+      "grad_norm": 0.302734375,
+      "learning_rate": 1.5189346650020808e-06,
+      "loss": 0.2891,
+      "step": 1460
+    },
+    {
+      "epoch": 0.6158967956720766,
+      "grad_norm": 0.291015625,
+      "learning_rate": 1.5397419891801914e-06,
+      "loss": 0.2913,
+      "step": 1480
+    },
+    {
+      "epoch": 0.6242197253433208,
+      "grad_norm": 0.3125,
+      "learning_rate": 1.5605493133583021e-06,
+      "loss": 0.2864,
+      "step": 1500
+    },
+    {
+      "epoch": 0.6242197253433208,
+      "eval_main_loss": 0.29658398032188416,
+      "eval_main_runtime": 6.3423,
+      "eval_main_samples_per_second": 29.958,
+      "eval_main_steps_per_second": 3.784,
+      "step": 1500
+    },
+    {
+      "epoch": 0.6242197253433208,
+      "eval_anatomy_loss": 2.939924478530884,
+      "eval_anatomy_runtime": 0.2668,
+      "eval_anatomy_samples_per_second": 7.496,
+      "eval_anatomy_steps_per_second": 3.748,
+      "step": 1500
+    },
+    {
+      "epoch": 0.6242197253433208,
+      "eval_college_mathematics_loss": 2.138369560241699,
+      "eval_college_mathematics_runtime": 0.2665,
+      "eval_college_mathematics_samples_per_second": 7.505,
+      "eval_college_mathematics_steps_per_second": 3.753,
+      "step": 1500
+    },
+    {
+      "epoch": 0.6242197253433208,
+      "eval_international_law_loss": 3.16491436958313,
+      "eval_international_law_runtime": 0.2662,
+      "eval_international_law_samples_per_second": 7.512,
+      "eval_international_law_steps_per_second": 3.756,
+      "step": 1500
+    },
+    {
+      "epoch": 0.6325426550145651,
+      "grad_norm": 0.294921875,
+      "learning_rate": 1.5813566375364128e-06,
+      "loss": 0.2904,
+      "step": 1520
+    },
+    {
+      "epoch": 0.6408655846858095,
+      "grad_norm": 0.294921875,
+      "learning_rate": 1.6021639617145237e-06,
+      "loss": 0.291,
+      "step": 1540
+    },
+    {
+      "epoch": 0.6491885143570537,
+      "grad_norm": 0.3203125,
+      "learning_rate": 1.6229712858926343e-06,
+      "loss": 0.287,
+      "step": 1560
+    },
+    {
+      "epoch": 0.657511444028298,
+      "grad_norm": 0.369140625,
+      "learning_rate": 1.643778610070745e-06,
+      "loss": 0.2892,
+      "step": 1580
+    },
+    {
+      "epoch": 0.6658343736995422,
+      "grad_norm": 0.328125,
+      "learning_rate": 1.6645859342488557e-06,
+      "loss": 0.2882,
+      "step": 1600
+    },
+    {
+      "epoch": 0.6658343736995422,
+      "eval_main_loss": 0.29542383551597595,
+      "eval_main_runtime": 6.3441,
+      "eval_main_samples_per_second": 29.949,
+      "eval_main_steps_per_second": 3.783,
+      "step": 1600
+    },
+    {
+      "epoch": 0.6658343736995422,
+      "eval_anatomy_loss": 2.932673692703247,
+      "eval_anatomy_runtime": 0.2664,
+      "eval_anatomy_samples_per_second": 7.508,
+      "eval_anatomy_steps_per_second": 3.754,
+      "step": 1600
+    },
+    {
+      "epoch": 0.6658343736995422,
+      "eval_college_mathematics_loss": 2.137777328491211,
+      "eval_college_mathematics_runtime": 0.2665,
+      "eval_college_mathematics_samples_per_second": 7.505,
+      "eval_college_mathematics_steps_per_second": 3.753,
+      "step": 1600
+    },
+    {
+      "epoch": 0.6658343736995422,
+      "eval_international_law_loss": 3.161029577255249,
+      "eval_international_law_runtime": 0.2659,
+      "eval_international_law_samples_per_second": 7.523,
+      "eval_international_law_steps_per_second": 3.761,
+      "step": 1600
+    },
+    {
+      "epoch": 0.6741573033707865,
+      "grad_norm": 0.302734375,
+      "learning_rate": 1.6853932584269663e-06,
+      "loss": 0.2896,
+      "step": 1620
+    },
+    {
+      "epoch": 0.6824802330420308,
+      "grad_norm": 0.349609375,
+      "learning_rate": 1.706200582605077e-06,
+      "loss": 0.29,
+      "step": 1640
+    },
+    {
+      "epoch": 0.690803162713275,
+      "grad_norm": 0.3671875,
+      "learning_rate": 1.7270079067831877e-06,
+      "loss": 0.2867,
+      "step": 1660
+    },
+    {
+      "epoch": 0.6991260923845194,
+      "grad_norm": 0.3671875,
+      "learning_rate": 1.7478152309612985e-06,
+      "loss": 0.2883,
+      "step": 1680
+    },
+    {
+      "epoch": 0.7074490220557637,
+      "grad_norm": 0.341796875,
+      "learning_rate": 1.7686225551394092e-06,
+      "loss": 0.2843,
+      "step": 1700
+    },
+    {
+      "epoch": 0.7074490220557637,
+      "eval_main_loss": 0.29412609338760376,
+      "eval_main_runtime": 6.3391,
+      "eval_main_samples_per_second": 29.973,
+      "eval_main_steps_per_second": 3.786,
+      "step": 1700
+    },
+    {
+      "epoch": 0.7074490220557637,
+      "eval_anatomy_loss": 2.92899751663208,
+      "eval_anatomy_runtime": 0.267,
+      "eval_anatomy_samples_per_second": 7.49,
+      "eval_anatomy_steps_per_second": 3.745,
+      "step": 1700
+    },
+    {
+      "epoch": 0.7074490220557637,
+      "eval_college_mathematics_loss": 2.132533311843872,
+      "eval_college_mathematics_runtime": 0.2655,
+      "eval_college_mathematics_samples_per_second": 7.532,
+      "eval_college_mathematics_steps_per_second": 3.766,
+      "step": 1700
+    },
+    {
+      "epoch": 0.7074490220557637,
+      "eval_international_law_loss": 3.1562838554382324,
+      "eval_international_law_runtime": 0.2661,
+      "eval_international_law_samples_per_second": 7.516,
+      "eval_international_law_steps_per_second": 3.758,
+      "step": 1700
+    },
+    {
+      "epoch": 0.7157719517270079,
+      "grad_norm": 0.373046875,
+      "learning_rate": 1.7894298793175199e-06,
+      "loss": 0.2916,
+      "step": 1720
+    },
+    {
+      "epoch": 0.7240948813982522,
+      "grad_norm": 0.373046875,
+      "learning_rate": 1.8102372034956305e-06,
+      "loss": 0.2844,
+      "step": 1740
+    },
+    {
+      "epoch": 0.7324178110694964,
+      "grad_norm": 0.408203125,
+      "learning_rate": 1.8310445276737412e-06,
+      "loss": 0.2877,
+      "step": 1760
+    },
+    {
+      "epoch": 0.7407407407407407,
+      "grad_norm": 0.3984375,
+      "learning_rate": 1.8518518518518519e-06,
+      "loss": 0.2862,
+      "step": 1780
+    },
+    {
+      "epoch": 0.7490636704119851,
+      "grad_norm": 0.421875,
+      "learning_rate": 1.8726591760299627e-06,
+      "loss": 0.2826,
+      "step": 1800
+    },
+    {
+      "epoch": 0.7490636704119851,
+      "eval_main_loss": 0.2929327189922333,
+      "eval_main_runtime": 6.3377,
+      "eval_main_samples_per_second": 29.979,
+      "eval_main_steps_per_second": 3.787,
+      "step": 1800
+    },
+    {
+      "epoch": 0.7490636704119851,
+      "eval_anatomy_loss": 2.923774480819702,
+      "eval_anatomy_runtime": 0.2657,
+      "eval_anatomy_samples_per_second": 7.527,
+      "eval_anatomy_steps_per_second": 3.763,
+      "step": 1800
+    },
+    {
+      "epoch": 0.7490636704119851,
+      "eval_college_mathematics_loss": 2.1297199726104736,
+      "eval_college_mathematics_runtime": 0.2661,
+      "eval_college_mathematics_samples_per_second": 7.516,
+      "eval_college_mathematics_steps_per_second": 3.758,
+      "step": 1800
+    },
+    {
+      "epoch": 0.7490636704119851,
+      "eval_international_law_loss": 3.153568744659424,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.505,
+      "eval_international_law_steps_per_second": 3.753,
+      "step": 1800
+    },
+    {
+      "epoch": 0.7573866000832293,
+      "grad_norm": 0.36328125,
+      "learning_rate": 1.8934665002080734e-06,
+      "loss": 0.2883,
+      "step": 1820
+    },
+    {
+      "epoch": 0.7657095297544736,
+      "grad_norm": 0.4140625,
+      "learning_rate": 1.9142738243861843e-06,
+      "loss": 0.2827,
+      "step": 1840
+    },
+    {
+      "epoch": 0.7740324594257179,
+      "grad_norm": 0.45703125,
+      "learning_rate": 1.9350811485642947e-06,
+      "loss": 0.2892,
+      "step": 1860
+    },
+    {
+      "epoch": 0.7823553890969621,
+      "grad_norm": 0.380859375,
+      "learning_rate": 1.9558884727424056e-06,
+      "loss": 0.2816,
+      "step": 1880
+    },
+    {
+      "epoch": 0.7906783187682064,
+      "grad_norm": 0.3671875,
+      "learning_rate": 1.976695796920516e-06,
+      "loss": 0.2859,
+      "step": 1900
+    },
+    {
+      "epoch": 0.7906783187682064,
+      "eval_main_loss": 0.2918354272842407,
+      "eval_main_runtime": 6.3351,
+      "eval_main_samples_per_second": 29.992,
+      "eval_main_steps_per_second": 3.788,
+      "step": 1900
+    },
+    {
+      "epoch": 0.7906783187682064,
+      "eval_anatomy_loss": 2.922013759613037,
+      "eval_anatomy_runtime": 0.2669,
+      "eval_anatomy_samples_per_second": 7.495,
+      "eval_anatomy_steps_per_second": 3.747,
+      "step": 1900
+    },
+    {
+      "epoch": 0.7906783187682064,
+      "eval_college_mathematics_loss": 2.127079963684082,
+      "eval_college_mathematics_runtime": 0.2667,
+      "eval_college_mathematics_samples_per_second": 7.498,
+      "eval_college_mathematics_steps_per_second": 3.749,
+      "step": 1900
+    },
+    {
+      "epoch": 0.7906783187682064,
+      "eval_international_law_loss": 3.1497292518615723,
+      "eval_international_law_runtime": 0.2681,
+      "eval_international_law_samples_per_second": 7.459,
+      "eval_international_law_steps_per_second": 3.729,
+      "step": 1900
+    },
+    {
+      "epoch": 0.7990012484394506,
+      "grad_norm": 0.44140625,
+      "learning_rate": 1.997503121098627e-06,
+      "loss": 0.2827,
+      "step": 1920
+    },
+    {
+      "epoch": 0.807324178110695,
+      "grad_norm": 0.396484375,
+      "learning_rate": 2.018310445276738e-06,
+      "loss": 0.2835,
+      "step": 1940
+    },
+    {
+      "epoch": 0.8156471077819393,
+      "grad_norm": 0.4375,
+      "learning_rate": 2.0391177694548483e-06,
+      "loss": 0.2824,
+      "step": 1960
+    },
+    {
+      "epoch": 0.8239700374531835,
+      "grad_norm": 0.4453125,
+      "learning_rate": 2.059925093632959e-06,
+      "loss": 0.2834,
+      "step": 1980
+    },
+    {
+      "epoch": 0.8322929671244278,
+      "grad_norm": 0.3984375,
+      "learning_rate": 2.0807324178110696e-06,
+      "loss": 0.285,
+      "step": 2000
+    },
+    {
+      "epoch": 0.8322929671244278,
+      "eval_main_loss": 0.29074886441230774,
+      "eval_main_runtime": 6.3397,
+      "eval_main_samples_per_second": 29.97,
+      "eval_main_steps_per_second": 3.786,
+      "step": 2000
+    },
+    {
+      "epoch": 0.8322929671244278,
+      "eval_anatomy_loss": 2.92093825340271,
+      "eval_anatomy_runtime": 0.2669,
+      "eval_anatomy_samples_per_second": 7.494,
+      "eval_anatomy_steps_per_second": 3.747,
+      "step": 2000
+    },
+    {
+      "epoch": 0.8322929671244278,
+      "eval_college_mathematics_loss": 2.1216940879821777,
+      "eval_college_mathematics_runtime": 0.2663,
+      "eval_college_mathematics_samples_per_second": 7.511,
+      "eval_college_mathematics_steps_per_second": 3.756,
+      "step": 2000
+    },
+    {
+      "epoch": 0.8322929671244278,
+      "eval_international_law_loss": 3.146960973739624,
+      "eval_international_law_runtime": 0.2679,
+      "eval_international_law_samples_per_second": 7.465,
+      "eval_international_law_steps_per_second": 3.733,
+      "step": 2000
+    },
+    {
+      "epoch": 0.8406158967956721,
+      "grad_norm": 0.41015625,
+      "learning_rate": 2.1015397419891805e-06,
+      "loss": 0.2793,
+      "step": 2020
+    },
+    {
+      "epoch": 0.8489388264669163,
+      "grad_norm": 0.443359375,
+      "learning_rate": 2.122347066167291e-06,
+      "loss": 0.2855,
+      "step": 2040
+    },
+    {
+      "epoch": 0.8572617561381606,
+      "grad_norm": 0.45703125,
+      "learning_rate": 2.1431543903454014e-06,
+      "loss": 0.2789,
+      "step": 2060
+    },
+    {
+      "epoch": 0.865584685809405,
+      "grad_norm": 0.4765625,
+      "learning_rate": 2.1639617145235127e-06,
+      "loss": 0.2807,
+      "step": 2080
+    },
+    {
+      "epoch": 0.8739076154806492,
+      "grad_norm": 0.431640625,
+      "learning_rate": 2.184769038701623e-06,
+      "loss": 0.2851,
+      "step": 2100
+    },
+    {
+      "epoch": 0.8739076154806492,
+      "eval_main_loss": 0.2896941602230072,
+      "eval_main_runtime": 6.3421,
+      "eval_main_samples_per_second": 29.958,
+      "eval_main_steps_per_second": 3.784,
+      "step": 2100
+    },
+    {
+      "epoch": 0.8739076154806492,
+      "eval_anatomy_loss": 2.916968822479248,
+      "eval_anatomy_runtime": 0.2673,
+      "eval_anatomy_samples_per_second": 7.481,
+      "eval_anatomy_steps_per_second": 3.741,
+      "step": 2100
+    },
+    {
+      "epoch": 0.8739076154806492,
+      "eval_college_mathematics_loss": 2.119070053100586,
+      "eval_college_mathematics_runtime": 0.2667,
+      "eval_college_mathematics_samples_per_second": 7.498,
+      "eval_college_mathematics_steps_per_second": 3.749,
+      "step": 2100
+    },
+    {
+      "epoch": 0.8739076154806492,
+      "eval_international_law_loss": 3.1451635360717773,
+      "eval_international_law_runtime": 0.2673,
+      "eval_international_law_samples_per_second": 7.481,
+      "eval_international_law_steps_per_second": 3.74,
+      "step": 2100
+    },
+    {
+      "epoch": 0.8822305451518935,
+      "grad_norm": 0.455078125,
+      "learning_rate": 2.205576362879734e-06,
+      "loss": 0.2894,
+      "step": 2120
+    },
+    {
+      "epoch": 0.8905534748231377,
+      "grad_norm": 0.431640625,
+      "learning_rate": 2.2263836870578445e-06,
+      "loss": 0.2872,
+      "step": 2140
+    },
+    {
+      "epoch": 0.898876404494382,
+      "grad_norm": 0.455078125,
+      "learning_rate": 2.2471910112359554e-06,
+      "loss": 0.2842,
+      "step": 2160
+    },
+    {
+      "epoch": 0.9071993341656263,
+      "grad_norm": 0.42578125,
+      "learning_rate": 2.267998335414066e-06,
+      "loss": 0.282,
+      "step": 2180
+    },
+    {
+      "epoch": 0.9155222638368706,
+      "grad_norm": 0.435546875,
+      "learning_rate": 2.2888056595921767e-06,
+      "loss": 0.2787,
+      "step": 2200
+    },
+    {
+      "epoch": 0.9155222638368706,
+      "eval_main_loss": 0.28874531388282776,
+      "eval_main_runtime": 6.3297,
+      "eval_main_samples_per_second": 30.017,
+      "eval_main_steps_per_second": 3.792,
+      "step": 2200
+    },
+    {
+      "epoch": 0.9155222638368706,
+      "eval_anatomy_loss": 2.9137964248657227,
+      "eval_anatomy_runtime": 0.2659,
+      "eval_anatomy_samples_per_second": 7.522,
+      "eval_anatomy_steps_per_second": 3.761,
+      "step": 2200
+    },
+    {
+      "epoch": 0.9155222638368706,
+      "eval_college_mathematics_loss": 2.115910530090332,
+      "eval_college_mathematics_runtime": 0.267,
+      "eval_college_mathematics_samples_per_second": 7.492,
+      "eval_college_mathematics_steps_per_second": 3.746,
+      "step": 2200
+    },
+    {
+      "epoch": 0.9155222638368706,
+      "eval_international_law_loss": 3.14198637008667,
+      "eval_international_law_runtime": 0.2668,
+      "eval_international_law_samples_per_second": 7.497,
+      "eval_international_law_steps_per_second": 3.749,
+      "step": 2200
+    },
+    {
+      "epoch": 0.9238451935081149,
+      "grad_norm": 0.474609375,
+      "learning_rate": 2.3096129837702876e-06,
+      "loss": 0.2844,
+      "step": 2220
+    },
+    {
+      "epoch": 0.9321681231793592,
+      "grad_norm": 0.44921875,
+      "learning_rate": 2.330420307948398e-06,
+      "loss": 0.2822,
+      "step": 2240
+    },
+    {
+      "epoch": 0.9404910528506034,
+      "grad_norm": 0.466796875,
+      "learning_rate": 2.351227632126509e-06,
+      "loss": 0.2776,
+      "step": 2260
+    },
+    {
+      "epoch": 0.9488139825218477,
+      "grad_norm": 0.388671875,
+      "learning_rate": 2.3720349563046194e-06,
+      "loss": 0.2788,
+      "step": 2280
+    },
+    {
+      "epoch": 0.9571369121930919,
+      "grad_norm": 0.486328125,
+      "learning_rate": 2.39284228048273e-06,
+      "loss": 0.2806,
+      "step": 2300
+    },
+    {
+      "epoch": 0.9571369121930919,
+      "eval_main_loss": 0.2877212464809418,
+      "eval_main_runtime": 6.3349,
+      "eval_main_samples_per_second": 29.993,
+      "eval_main_steps_per_second": 3.789,
+      "step": 2300
+    },
+    {
+      "epoch": 0.9571369121930919,
+      "eval_anatomy_loss": 2.9125289916992188,
+      "eval_anatomy_runtime": 0.2664,
+      "eval_anatomy_samples_per_second": 7.507,
+      "eval_anatomy_steps_per_second": 3.753,
+      "step": 2300
+    },
+    {
+      "epoch": 0.9571369121930919,
+      "eval_college_mathematics_loss": 2.1152408123016357,
+      "eval_college_mathematics_runtime": 0.2681,
+      "eval_college_mathematics_samples_per_second": 7.459,
+      "eval_college_mathematics_steps_per_second": 3.729,
+      "step": 2300
+    },
+    {
+      "epoch": 0.9571369121930919,
+      "eval_international_law_loss": 3.1422488689422607,
+      "eval_international_law_runtime": 0.2661,
+      "eval_international_law_samples_per_second": 7.516,
+      "eval_international_law_steps_per_second": 3.758,
+      "step": 2300
+    },
+    {
+      "epoch": 0.9654598418643362,
+      "grad_norm": 0.490234375,
+      "learning_rate": 2.4136496046608407e-06,
+      "loss": 0.2781,
+      "step": 2320
+    },
+    {
+      "epoch": 0.9737827715355806,
+      "grad_norm": 0.4296875,
+      "learning_rate": 2.4344569288389516e-06,
+      "loss": 0.2808,
+      "step": 2340
+    },
+    {
+      "epoch": 0.9821057012068248,
+      "grad_norm": 0.4609375,
+      "learning_rate": 2.4552642530170625e-06,
+      "loss": 0.2829,
+      "step": 2360
+    },
+    {
+      "epoch": 0.9904286308780691,
+      "grad_norm": 0.466796875,
+      "learning_rate": 2.476071577195173e-06,
+      "loss": 0.2789,
+      "step": 2380
+    },
+    {
+      "epoch": 0.9987515605493134,
+      "grad_norm": 0.44921875,
+      "learning_rate": 2.4968789013732834e-06,
+      "loss": 0.283,
+      "step": 2400
+    },
+    {
+      "epoch": 0.9987515605493134,
+      "eval_main_loss": 0.2861831784248352,
+      "eval_main_runtime": 6.3371,
+      "eval_main_samples_per_second": 29.982,
+      "eval_main_steps_per_second": 3.787,
+      "step": 2400
+    },
+    {
+      "epoch": 0.9987515605493134,
+      "eval_anatomy_loss": 2.9071733951568604,
+      "eval_anatomy_runtime": 0.2662,
+      "eval_anatomy_samples_per_second": 7.513,
+      "eval_anatomy_steps_per_second": 3.756,
+      "step": 2400
+    },
+    {
+      "epoch": 0.9987515605493134,
+      "eval_college_mathematics_loss": 2.1126503944396973,
+      "eval_college_mathematics_runtime": 0.2665,
+      "eval_college_mathematics_samples_per_second": 7.503,
+      "eval_college_mathematics_steps_per_second": 3.752,
+      "step": 2400
+    },
+    {
+      "epoch": 0.9987515605493134,
+      "eval_international_law_loss": 3.1375622749328613,
+      "eval_international_law_runtime": 0.267,
+      "eval_international_law_samples_per_second": 7.492,
+      "eval_international_law_steps_per_second": 3.746,
+      "step": 2400
+    },
+    {
+      "epoch": 1.0070744902205577,
+      "grad_norm": 0.474609375,
+      "learning_rate": 2.5176862255513947e-06,
+      "loss": 0.2779,
+      "step": 2420
+    },
+    {
+      "epoch": 1.0153974198918019,
+      "grad_norm": 0.51171875,
+      "learning_rate": 2.5384935497295047e-06,
+      "loss": 0.282,
+      "step": 2440
+    },
+    {
+      "epoch": 1.0237203495630463,
+      "grad_norm": 0.435546875,
+      "learning_rate": 2.559300873907616e-06,
+      "loss": 0.2757,
+      "step": 2460
+    },
+    {
+      "epoch": 1.0320432792342904,
+      "grad_norm": 0.431640625,
+      "learning_rate": 2.580108198085726e-06,
+      "loss": 0.2803,
+      "step": 2480
+    },
+    {
+      "epoch": 1.0403662089055348,
+      "grad_norm": 0.46484375,
+      "learning_rate": 2.6009155222638374e-06,
+      "loss": 0.2768,
+      "step": 2500
+    },
+    {
+      "epoch": 1.0403662089055348,
+      "eval_main_loss": 0.283886194229126,
+      "eval_main_runtime": 6.3349,
+      "eval_main_samples_per_second": 29.992,
+      "eval_main_steps_per_second": 3.789,
+      "step": 2500
+    },
+    {
+      "epoch": 1.0403662089055348,
+      "eval_anatomy_loss": 2.9029533863067627,
+      "eval_anatomy_runtime": 0.266,
+      "eval_anatomy_samples_per_second": 7.519,
+      "eval_anatomy_steps_per_second": 3.76,
+      "step": 2500
+    },
+    {
+      "epoch": 1.0403662089055348,
+      "eval_college_mathematics_loss": 2.1050500869750977,
+      "eval_college_mathematics_runtime": 0.2663,
+      "eval_college_mathematics_samples_per_second": 7.51,
+      "eval_college_mathematics_steps_per_second": 3.755,
+      "step": 2500
+    },
+    {
+      "epoch": 1.0403662089055348,
+      "eval_international_law_loss": 3.1294405460357666,
+      "eval_international_law_runtime": 0.2672,
+      "eval_international_law_samples_per_second": 7.484,
+      "eval_international_law_steps_per_second": 3.742,
+      "step": 2500
+    },
+    {
+      "epoch": 1.048689138576779,
+      "grad_norm": 0.462890625,
+      "learning_rate": 2.6217228464419474e-06,
+      "loss": 0.2817,
+      "step": 2520
+    },
+    {
+      "epoch": 1.0570120682480233,
+      "grad_norm": 0.5078125,
+      "learning_rate": 2.6425301706200583e-06,
+      "loss": 0.2763,
+      "step": 2540
+    },
+    {
+      "epoch": 1.0653349979192677,
+      "grad_norm": 0.46484375,
+      "learning_rate": 2.6633374947981696e-06,
+      "loss": 0.2757,
+      "step": 2560
+    },
+    {
+      "epoch": 1.0736579275905118,
+      "grad_norm": 0.51953125,
+      "learning_rate": 2.6841448189762796e-06,
+      "loss": 0.2764,
+      "step": 2580
+    },
+    {
+      "epoch": 1.0819808572617562,
+      "grad_norm": 0.48828125,
+      "learning_rate": 2.704952143154391e-06,
+      "loss": 0.2747,
+      "step": 2600
+    },
+    {
+      "epoch": 1.0819808572617562,
+      "eval_main_loss": 0.2814878225326538,
+      "eval_main_runtime": 6.3334,
+      "eval_main_samples_per_second": 30.0,
+      "eval_main_steps_per_second": 3.789,
+      "step": 2600
+    },
+    {
+      "epoch": 1.0819808572617562,
+      "eval_anatomy_loss": 2.898045778274536,
+      "eval_anatomy_runtime": 0.2665,
+      "eval_anatomy_samples_per_second": 7.503,
+      "eval_anatomy_steps_per_second": 3.752,
+      "step": 2600
+    },
+    {
+      "epoch": 1.0819808572617562,
+      "eval_college_mathematics_loss": 2.1027755737304688,
+      "eval_college_mathematics_runtime": 0.2664,
+      "eval_college_mathematics_samples_per_second": 7.509,
+      "eval_college_mathematics_steps_per_second": 3.754,
+      "step": 2600
+    },
+    {
+      "epoch": 1.0819808572617562,
+      "eval_international_law_loss": 3.1270911693573,
+      "eval_international_law_runtime": 0.2666,
+      "eval_international_law_samples_per_second": 7.503,
+      "eval_international_law_steps_per_second": 3.751,
+      "step": 2600
+    },
+    {
+      "epoch": 1.0903037869330003,
+      "grad_norm": 0.474609375,
+      "learning_rate": 2.725759467332501e-06,
+      "loss": 0.2752,
+      "step": 2620
+    },
+    {
+      "epoch": 1.0986267166042447,
+      "grad_norm": 0.49609375,
+      "learning_rate": 2.746566791510612e-06,
+      "loss": 0.2709,
+      "step": 2640
+    },
+    {
+      "epoch": 1.1069496462754889,
+      "grad_norm": 0.490234375,
+      "learning_rate": 2.7673741156887223e-06,
+      "loss": 0.2741,
+      "step": 2660
+    },
+    {
+      "epoch": 1.1152725759467332,
+      "grad_norm": 0.44140625,
+      "learning_rate": 2.788181439866833e-06,
+      "loss": 0.2693,
+      "step": 2680
+    },
+    {
+      "epoch": 1.1235955056179776,
+      "grad_norm": 0.474609375,
+      "learning_rate": 2.8089887640449444e-06,
+      "loss": 0.2796,
+      "step": 2700
+    },
+    {
+      "epoch": 1.1235955056179776,
+      "eval_main_loss": 0.2786170542240143,
+      "eval_main_runtime": 6.3383,
+      "eval_main_samples_per_second": 29.977,
+      "eval_main_steps_per_second": 3.787,
+      "step": 2700
+    },
+    {
+      "epoch": 1.1235955056179776,
+      "eval_anatomy_loss": 2.8903613090515137,
+      "eval_anatomy_runtime": 0.2661,
+      "eval_anatomy_samples_per_second": 7.517,
+      "eval_anatomy_steps_per_second": 3.758,
+      "step": 2700
+    },
+    {
+      "epoch": 1.1235955056179776,
+      "eval_college_mathematics_loss": 2.0952742099761963,
+      "eval_college_mathematics_runtime": 0.2653,
+      "eval_college_mathematics_samples_per_second": 7.537,
+      "eval_college_mathematics_steps_per_second": 3.769,
+      "step": 2700
+    },
+    {
+      "epoch": 1.1235955056179776,
+      "eval_international_law_loss": 3.120177745819092,
+      "eval_international_law_runtime": 0.2677,
+      "eval_international_law_samples_per_second": 7.47,
+      "eval_international_law_steps_per_second": 3.735,
+      "step": 2700
+    },
+    {
+      "epoch": 1.1319184352892218,
+      "grad_norm": 0.50390625,
+      "learning_rate": 2.8297960882230545e-06,
+      "loss": 0.2717,
+      "step": 2720
+    },
+    {
+      "epoch": 1.1402413649604661,
+      "grad_norm": 0.498046875,
+      "learning_rate": 2.8506034124011653e-06,
+      "loss": 0.2704,
+      "step": 2740
+    },
+    {
+      "epoch": 1.1485642946317103,
+      "grad_norm": 0.4765625,
+      "learning_rate": 2.871410736579276e-06,
+      "loss": 0.2688,
+      "step": 2760
+    },
+    {
+      "epoch": 1.1568872243029547,
+      "grad_norm": 0.458984375,
+      "learning_rate": 2.8922180607573867e-06,
+      "loss": 0.2654,
+      "step": 2780
+    },
+    {
+      "epoch": 1.1652101539741988,
+      "grad_norm": 0.419921875,
+      "learning_rate": 2.913025384935498e-06,
+      "loss": 0.2646,
+      "step": 2800
+    },
+    {
+      "epoch": 1.1652101539741988,
+      "eval_main_loss": 0.27353334426879883,
+      "eval_main_runtime": 6.3411,
+      "eval_main_samples_per_second": 29.963,
+      "eval_main_steps_per_second": 3.785,
+      "step": 2800
+    },
+    {
+      "epoch": 1.1652101539741988,
+      "eval_anatomy_loss": 2.8842265605926514,
+      "eval_anatomy_runtime": 0.2659,
+      "eval_anatomy_samples_per_second": 7.521,
+      "eval_anatomy_steps_per_second": 3.76,
+      "step": 2800
+    },
+    {
+      "epoch": 1.1652101539741988,
+      "eval_college_mathematics_loss": 2.086782693862915,
+      "eval_college_mathematics_runtime": 0.2656,
+      "eval_college_mathematics_samples_per_second": 7.531,
+      "eval_college_mathematics_steps_per_second": 3.766,
+      "step": 2800
+    },
+    {
+      "epoch": 1.1652101539741988,
+      "eval_international_law_loss": 3.11441707611084,
+      "eval_international_law_runtime": 0.2666,
+      "eval_international_law_samples_per_second": 7.501,
+      "eval_international_law_steps_per_second": 3.75,
+      "step": 2800
+    },
+    {
+      "epoch": 1.1735330836454432,
+      "grad_norm": 0.43359375,
+      "learning_rate": 2.933832709113608e-06,
+      "loss": 0.2618,
+      "step": 2820
+    },
+    {
+      "epoch": 1.1818560133166875,
+      "grad_norm": 0.453125,
+      "learning_rate": 2.9546400332917193e-06,
+      "loss": 0.2636,
+      "step": 2840
+    },
+    {
+      "epoch": 1.1901789429879317,
+      "grad_norm": 0.453125,
+      "learning_rate": 2.9754473574698293e-06,
+      "loss": 0.2654,
+      "step": 2860
+    },
+    {
+      "epoch": 1.198501872659176,
+      "grad_norm": 0.46484375,
+      "learning_rate": 2.9962546816479402e-06,
+      "loss": 0.2586,
+      "step": 2880
+    },
+    {
+      "epoch": 1.2068248023304202,
+      "grad_norm": 0.462890625,
+      "learning_rate": 3.0170620058260507e-06,
+      "loss": 0.2576,
+      "step": 2900
+    },
+    {
+      "epoch": 1.2068248023304202,
+      "eval_main_loss": 0.2663625478744507,
+      "eval_main_runtime": 6.3386,
+      "eval_main_samples_per_second": 29.975,
+      "eval_main_steps_per_second": 3.786,
+      "step": 2900
+    },
+    {
+      "epoch": 1.2068248023304202,
+      "eval_anatomy_loss": 2.876462936401367,
+      "eval_anatomy_runtime": 0.2665,
+      "eval_anatomy_samples_per_second": 7.503,
+      "eval_anatomy_steps_per_second": 3.752,
+      "step": 2900
+    },
+    {
+      "epoch": 1.2068248023304202,
+      "eval_college_mathematics_loss": 2.0814149379730225,
+      "eval_college_mathematics_runtime": 0.2676,
+      "eval_college_mathematics_samples_per_second": 7.473,
+      "eval_college_mathematics_steps_per_second": 3.736,
+      "step": 2900
+    },
+    {
+      "epoch": 1.2068248023304202,
+      "eval_international_law_loss": 3.108351707458496,
+      "eval_international_law_runtime": 0.2657,
+      "eval_international_law_samples_per_second": 7.527,
+      "eval_international_law_steps_per_second": 3.764,
+      "step": 2900
+    },
+    {
+      "epoch": 1.2151477320016646,
+      "grad_norm": 0.41015625,
+      "learning_rate": 3.0378693300041616e-06,
+      "loss": 0.2599,
+      "step": 2920
+    },
+    {
+      "epoch": 1.2234706616729087,
+      "grad_norm": 0.46875,
+      "learning_rate": 3.058676654182273e-06,
+      "loss": 0.2602,
+      "step": 2940
+    },
+    {
+      "epoch": 1.2317935913441531,
+      "grad_norm": 0.4609375,
+      "learning_rate": 3.079483978360383e-06,
+      "loss": 0.2564,
+      "step": 2960
+    },
+    {
+      "epoch": 1.2401165210153975,
+      "grad_norm": 0.42578125,
+      "learning_rate": 3.1002913025384938e-06,
+      "loss": 0.2551,
+      "step": 2980
+    },
+    {
+      "epoch": 1.2484394506866416,
+      "grad_norm": 0.416015625,
+      "learning_rate": 3.1210986267166042e-06,
+      "loss": 0.2551,
+      "step": 3000
+    },
+    {
+      "epoch": 1.2484394506866416,
+      "eval_main_loss": 0.2596937417984009,
+      "eval_main_runtime": 6.3395,
+      "eval_main_samples_per_second": 29.971,
+      "eval_main_steps_per_second": 3.786,
+      "step": 3000
+    },
+    {
+      "epoch": 1.2484394506866416,
+      "eval_anatomy_loss": 2.8675546646118164,
+      "eval_anatomy_runtime": 0.2666,
+      "eval_anatomy_samples_per_second": 7.501,
+      "eval_anatomy_steps_per_second": 3.751,
+      "step": 3000
+    },
+    {
+      "epoch": 1.2484394506866416,
+      "eval_college_mathematics_loss": 2.0764455795288086,
+      "eval_college_mathematics_runtime": 0.2661,
+      "eval_college_mathematics_samples_per_second": 7.515,
+      "eval_college_mathematics_steps_per_second": 3.758,
+      "step": 3000
+    },
+    {
+      "epoch": 1.2484394506866416,
+      "eval_international_law_loss": 3.1014602184295654,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.505,
+      "eval_international_law_steps_per_second": 3.752,
+      "step": 3000
+    },
+    {
+      "epoch": 1.256762380357886,
+      "grad_norm": 0.423828125,
+      "learning_rate": 3.141905950894715e-06,
+      "loss": 0.2528,
+      "step": 3020
+    },
+    {
+      "epoch": 1.2650853100291304,
+      "grad_norm": 0.408203125,
+      "learning_rate": 3.1627132750728256e-06,
+      "loss": 0.2555,
+      "step": 3040
+    },
+    {
+      "epoch": 1.2734082397003745,
+      "grad_norm": 0.404296875,
+      "learning_rate": 3.1835205992509364e-06,
+      "loss": 0.2527,
+      "step": 3060
+    },
+    {
+      "epoch": 1.2817311693716187,
+      "grad_norm": 0.41796875,
+      "learning_rate": 3.2043279234290473e-06,
+      "loss": 0.2523,
+      "step": 3080
+    },
+    {
+      "epoch": 1.290054099042863,
+      "grad_norm": 0.4453125,
+      "learning_rate": 3.2251352476071578e-06,
+      "loss": 0.2501,
+      "step": 3100
+    },
+    {
+      "epoch": 1.290054099042863,
+      "eval_main_loss": 0.2549309730529785,
+      "eval_main_runtime": 6.3572,
+      "eval_main_samples_per_second": 29.887,
+      "eval_main_steps_per_second": 3.775,
+      "step": 3100
+    },
+    {
+      "epoch": 1.290054099042863,
+      "eval_anatomy_loss": 2.862455368041992,
+      "eval_anatomy_runtime": 0.2664,
+      "eval_anatomy_samples_per_second": 7.507,
+      "eval_anatomy_steps_per_second": 3.753,
+      "step": 3100
+    },
+    {
+      "epoch": 1.290054099042863,
+      "eval_college_mathematics_loss": 2.0720179080963135,
+      "eval_college_mathematics_runtime": 0.2673,
+      "eval_college_mathematics_samples_per_second": 7.482,
+      "eval_college_mathematics_steps_per_second": 3.741,
+      "step": 3100
+    },
+    {
+      "epoch": 1.290054099042863,
+      "eval_international_law_loss": 3.095749855041504,
+      "eval_international_law_runtime": 0.2682,
+      "eval_international_law_samples_per_second": 7.457,
+      "eval_international_law_steps_per_second": 3.728,
+      "step": 3100
+    },
+    {
+      "epoch": 1.2983770287141074,
+      "grad_norm": 0.423828125,
+      "learning_rate": 3.2459425717852687e-06,
+      "loss": 0.2479,
+      "step": 3120
+    },
+    {
+      "epoch": 1.3066999583853516,
+      "grad_norm": 0.376953125,
+      "learning_rate": 3.266749895963379e-06,
+      "loss": 0.2454,
+      "step": 3140
+    },
+    {
+      "epoch": 1.315022888056596,
+      "grad_norm": 0.41796875,
+      "learning_rate": 3.28755722014149e-06,
+      "loss": 0.2428,
+      "step": 3160
+    },
+    {
+      "epoch": 1.3233458177278403,
+      "grad_norm": 0.404296875,
+      "learning_rate": 3.3083645443196004e-06,
+      "loss": 0.243,
+      "step": 3180
+    },
+    {
+      "epoch": 1.3316687473990845,
+      "grad_norm": 0.38671875,
+      "learning_rate": 3.3291718684977113e-06,
+      "loss": 0.2489,
+      "step": 3200
+    },
+    {
+      "epoch": 1.3316687473990845,
+      "eval_main_loss": 0.2516440749168396,
+      "eval_main_runtime": 6.3553,
+      "eval_main_samples_per_second": 29.896,
+      "eval_main_steps_per_second": 3.776,
+      "step": 3200
+    },
+    {
+      "epoch": 1.3316687473990845,
+      "eval_anatomy_loss": 2.8593392372131348,
+      "eval_anatomy_runtime": 0.2673,
+      "eval_anatomy_samples_per_second": 7.481,
+      "eval_anatomy_steps_per_second": 3.741,
+      "step": 3200
+    },
+    {
+      "epoch": 1.3316687473990845,
+      "eval_college_mathematics_loss": 2.0700876712799072,
+      "eval_college_mathematics_runtime": 0.2677,
+      "eval_college_mathematics_samples_per_second": 7.471,
+      "eval_college_mathematics_steps_per_second": 3.735,
+      "step": 3200
+    },
+    {
+      "epoch": 1.3316687473990845,
+      "eval_international_law_loss": 3.094395875930786,
+      "eval_international_law_runtime": 0.2675,
+      "eval_international_law_samples_per_second": 7.478,
+      "eval_international_law_steps_per_second": 3.739,
+      "step": 3200
+    },
+    {
+      "epoch": 1.3399916770703286,
+      "grad_norm": 0.34765625,
+      "learning_rate": 3.349979192675822e-06,
+      "loss": 0.2518,
+      "step": 3220
+    },
+    {
+      "epoch": 1.348314606741573,
+      "grad_norm": 0.412109375,
+      "learning_rate": 3.3707865168539327e-06,
+      "loss": 0.243,
+      "step": 3240
+    },
+    {
+      "epoch": 1.3566375364128174,
+      "grad_norm": 0.431640625,
+      "learning_rate": 3.3915938410320435e-06,
+      "loss": 0.246,
+      "step": 3260
+    },
+    {
+      "epoch": 1.3649604660840615,
+      "grad_norm": 0.322265625,
+      "learning_rate": 3.412401165210154e-06,
+      "loss": 0.249,
+      "step": 3280
+    },
+    {
+      "epoch": 1.373283395755306,
+      "grad_norm": 0.37890625,
+      "learning_rate": 3.433208489388265e-06,
+      "loss": 0.2439,
+      "step": 3300
+    },
+    {
+      "epoch": 1.373283395755306,
+      "eval_main_loss": 0.24985496699810028,
+      "eval_main_runtime": 6.3475,
+      "eval_main_samples_per_second": 29.933,
+      "eval_main_steps_per_second": 3.781,
+      "step": 3300
+    },
+    {
+      "epoch": 1.373283395755306,
+      "eval_anatomy_loss": 2.8591318130493164,
+      "eval_anatomy_runtime": 0.2662,
+      "eval_anatomy_samples_per_second": 7.514,
+      "eval_anatomy_steps_per_second": 3.757,
+      "step": 3300
+    },
+    {
+      "epoch": 1.373283395755306,
+      "eval_college_mathematics_loss": 2.067070722579956,
+      "eval_college_mathematics_runtime": 0.2676,
+      "eval_college_mathematics_samples_per_second": 7.475,
+      "eval_college_mathematics_steps_per_second": 3.738,
+      "step": 3300
+    },
+    {
+      "epoch": 1.373283395755306,
+      "eval_international_law_loss": 3.0916695594787598,
+      "eval_international_law_runtime": 0.2676,
+      "eval_international_law_samples_per_second": 7.475,
+      "eval_international_law_steps_per_second": 3.738,
+      "step": 3300
+    },
+    {
+      "epoch": 1.3816063254265503,
+      "grad_norm": 0.369140625,
+      "learning_rate": 3.4540158135663753e-06,
+      "loss": 0.2431,
+      "step": 3320
+    },
+    {
+      "epoch": 1.3899292550977944,
+      "grad_norm": 0.416015625,
+      "learning_rate": 3.474823137744486e-06,
+      "loss": 0.2437,
+      "step": 3340
+    },
+    {
+      "epoch": 1.3982521847690386,
+      "grad_norm": 0.353515625,
+      "learning_rate": 3.495630461922597e-06,
+      "loss": 0.2434,
+      "step": 3360
+    },
+    {
+      "epoch": 1.406575114440283,
+      "grad_norm": 0.376953125,
+      "learning_rate": 3.5164377861007075e-06,
+      "loss": 0.243,
+      "step": 3380
+    },
+    {
+      "epoch": 1.4148980441115273,
+      "grad_norm": 0.40625,
+      "learning_rate": 3.5372451102788184e-06,
+      "loss": 0.2447,
+      "step": 3400
+    },
+    {
+      "epoch": 1.4148980441115273,
+      "eval_main_loss": 0.24850161373615265,
+      "eval_main_runtime": 6.3478,
+      "eval_main_samples_per_second": 29.931,
+      "eval_main_steps_per_second": 3.781,
+      "step": 3400
+    },
+    {
+      "epoch": 1.4148980441115273,
+      "eval_anatomy_loss": 2.85494327545166,
+      "eval_anatomy_runtime": 0.2665,
+      "eval_anatomy_samples_per_second": 7.505,
+      "eval_anatomy_steps_per_second": 3.753,
+      "step": 3400
+    },
+    {
+      "epoch": 1.4148980441115273,
+      "eval_college_mathematics_loss": 2.071016550064087,
+      "eval_college_mathematics_runtime": 0.2668,
+      "eval_college_mathematics_samples_per_second": 7.496,
+      "eval_college_mathematics_steps_per_second": 3.748,
+      "step": 3400
+    },
+    {
+      "epoch": 1.4148980441115273,
+      "eval_international_law_loss": 3.0913243293762207,
+      "eval_international_law_runtime": 0.2682,
+      "eval_international_law_samples_per_second": 7.457,
+      "eval_international_law_steps_per_second": 3.729,
+      "step": 3400
+    },
+    {
+      "epoch": 1.4232209737827715,
+      "grad_norm": 0.3828125,
+      "learning_rate": 3.558052434456929e-06,
+      "loss": 0.2432,
+      "step": 3420
+    },
+    {
+      "epoch": 1.4315439034540158,
+      "grad_norm": 0.37890625,
+      "learning_rate": 3.5788597586350397e-06,
+      "loss": 0.2367,
+      "step": 3440
+    },
+    {
+      "epoch": 1.4398668331252602,
+      "grad_norm": 0.357421875,
+      "learning_rate": 3.59966708281315e-06,
+      "loss": 0.242,
+      "step": 3460
+    },
+    {
+      "epoch": 1.4481897627965044,
+      "grad_norm": 0.369140625,
+      "learning_rate": 3.620474406991261e-06,
+      "loss": 0.2382,
+      "step": 3480
+    },
+    {
+      "epoch": 1.4565126924677487,
+      "grad_norm": 0.34765625,
+      "learning_rate": 3.641281731169372e-06,
+      "loss": 0.2412,
+      "step": 3500
+    },
+    {
+      "epoch": 1.4565126924677487,
+      "eval_main_loss": 0.24757333099842072,
+      "eval_main_runtime": 6.3261,
+      "eval_main_samples_per_second": 30.034,
+      "eval_main_steps_per_second": 3.794,
+      "step": 3500
+    },
+    {
+      "epoch": 1.4565126924677487,
+      "eval_anatomy_loss": 2.8548152446746826,
+      "eval_anatomy_runtime": 0.2658,
+      "eval_anatomy_samples_per_second": 7.525,
+      "eval_anatomy_steps_per_second": 3.763,
+      "step": 3500
+    },
+    {
+      "epoch": 1.4565126924677487,
+      "eval_college_mathematics_loss": 2.0659983158111572,
+      "eval_college_mathematics_runtime": 0.2656,
+      "eval_college_mathematics_samples_per_second": 7.531,
+      "eval_college_mathematics_steps_per_second": 3.766,
+      "step": 3500
+    },
+    {
+      "epoch": 1.4565126924677487,
+      "eval_international_law_loss": 3.090118646621704,
+      "eval_international_law_runtime": 0.2671,
+      "eval_international_law_samples_per_second": 7.489,
+      "eval_international_law_steps_per_second": 3.745,
+      "step": 3500
+    },
+    {
+      "epoch": 1.4648356221389929,
+      "grad_norm": 0.375,
+      "learning_rate": 3.6620890553474824e-06,
+      "loss": 0.2393,
+      "step": 3520
+    },
+    {
+      "epoch": 1.4731585518102372,
+      "grad_norm": 0.365234375,
+      "learning_rate": 3.6828963795255933e-06,
+      "loss": 0.2408,
+      "step": 3540
+    },
+    {
+      "epoch": 1.4814814814814814,
+      "grad_norm": 0.322265625,
+      "learning_rate": 3.7037037037037037e-06,
+      "loss": 0.2421,
+      "step": 3560
+    },
+    {
+      "epoch": 1.4898044111527258,
+      "grad_norm": 0.333984375,
+      "learning_rate": 3.7245110278818146e-06,
+      "loss": 0.2396,
+      "step": 3580
+    },
+    {
+      "epoch": 1.4981273408239701,
+      "grad_norm": 0.390625,
+      "learning_rate": 3.7453183520599255e-06,
+      "loss": 0.2409,
+      "step": 3600
+    },
+    {
+      "epoch": 1.4981273408239701,
+      "eval_main_loss": 0.24677349627017975,
+      "eval_main_runtime": 6.3243,
+      "eval_main_samples_per_second": 30.043,
+      "eval_main_steps_per_second": 3.795,
+      "step": 3600
+    },
+    {
+      "epoch": 1.4981273408239701,
+      "eval_anatomy_loss": 2.8548572063446045,
+      "eval_anatomy_runtime": 0.2652,
+      "eval_anatomy_samples_per_second": 7.542,
+      "eval_anatomy_steps_per_second": 3.771,
+      "step": 3600
+    },
+    {
+      "epoch": 1.4981273408239701,
+      "eval_college_mathematics_loss": 2.067040205001831,
+      "eval_college_mathematics_runtime": 0.2656,
+      "eval_college_mathematics_samples_per_second": 7.529,
+      "eval_college_mathematics_steps_per_second": 3.764,
+      "step": 3600
+    },
+    {
+      "epoch": 1.4981273408239701,
+      "eval_international_law_loss": 3.08817720413208,
+      "eval_international_law_runtime": 0.2659,
+      "eval_international_law_samples_per_second": 7.523,
+      "eval_international_law_steps_per_second": 3.761,
+      "step": 3600
+    },
+    {
+      "epoch": 1.5064502704952143,
+      "grad_norm": 0.3828125,
+      "learning_rate": 3.766125676238036e-06,
+      "loss": 0.24,
+      "step": 3620
+    },
+    {
+      "epoch": 1.5147732001664584,
+      "grad_norm": 0.37109375,
+      "learning_rate": 3.786933000416147e-06,
+      "loss": 0.2414,
+      "step": 3640
+    },
+    {
+      "epoch": 1.5230961298377028,
+      "grad_norm": 0.37109375,
+      "learning_rate": 3.8077403245942573e-06,
+      "loss": 0.2401,
+      "step": 3660
+    },
+    {
+      "epoch": 1.5314190595089472,
+      "grad_norm": 0.37890625,
+      "learning_rate": 3.828547648772369e-06,
+      "loss": 0.2373,
+      "step": 3680
+    },
+    {
+      "epoch": 1.5397419891801913,
+      "grad_norm": 0.353515625,
+      "learning_rate": 3.849354972950479e-06,
+      "loss": 0.2404,
+      "step": 3700
+    },
+    {
+      "epoch": 1.5397419891801913,
+      "eval_main_loss": 0.24623610079288483,
+      "eval_main_runtime": 6.3272,
+      "eval_main_samples_per_second": 30.029,
+      "eval_main_steps_per_second": 3.793,
+      "step": 3700
+    },
+    {
+      "epoch": 1.5397419891801913,
+      "eval_anatomy_loss": 2.8530216217041016,
+      "eval_anatomy_runtime": 0.266,
+      "eval_anatomy_samples_per_second": 7.518,
+      "eval_anatomy_steps_per_second": 3.759,
+      "step": 3700
+    },
+    {
+      "epoch": 1.5397419891801913,
+      "eval_college_mathematics_loss": 2.065014123916626,
+      "eval_college_mathematics_runtime": 0.2657,
+      "eval_college_mathematics_samples_per_second": 7.528,
+      "eval_college_mathematics_steps_per_second": 3.764,
+      "step": 3700
+    },
+    {
+      "epoch": 1.5397419891801913,
+      "eval_international_law_loss": 3.0873329639434814,
+      "eval_international_law_runtime": 0.2663,
+      "eval_international_law_samples_per_second": 7.51,
+      "eval_international_law_steps_per_second": 3.755,
+      "step": 3700
+    },
+    {
+      "epoch": 1.5480649188514357,
+      "grad_norm": 0.34765625,
+      "learning_rate": 3.8701622971285895e-06,
+      "loss": 0.2414,
+      "step": 3720
+    },
+    {
+      "epoch": 1.55638784852268,
+      "grad_norm": 0.384765625,
+      "learning_rate": 3.890969621306701e-06,
+      "loss": 0.2402,
+      "step": 3740
+    },
+    {
+      "epoch": 1.5647107781939242,
+      "grad_norm": 0.400390625,
+      "learning_rate": 3.911776945484811e-06,
+      "loss": 0.2394,
+      "step": 3760
+    },
+    {
+      "epoch": 1.5730337078651684,
+      "grad_norm": 0.328125,
+      "learning_rate": 3.932584269662922e-06,
+      "loss": 0.239,
+      "step": 3780
+    },
+    {
+      "epoch": 1.581356637536413,
+      "grad_norm": 0.384765625,
+      "learning_rate": 3.953391593841032e-06,
+      "loss": 0.2383,
+      "step": 3800
+    },
+    {
+      "epoch": 1.581356637536413,
+      "eval_main_loss": 0.24581117928028107,
+      "eval_main_runtime": 6.3278,
+      "eval_main_samples_per_second": 30.026,
+      "eval_main_steps_per_second": 3.793,
+      "step": 3800
+    },
+    {
+      "epoch": 1.581356637536413,
+      "eval_anatomy_loss": 2.8533191680908203,
+      "eval_anatomy_runtime": 0.2681,
+      "eval_anatomy_samples_per_second": 7.459,
+      "eval_anatomy_steps_per_second": 3.729,
+      "step": 3800
+    },
+    {
+      "epoch": 1.581356637536413,
+      "eval_college_mathematics_loss": 2.0673060417175293,
+      "eval_college_mathematics_runtime": 0.2658,
+      "eval_college_mathematics_samples_per_second": 7.526,
+      "eval_college_mathematics_steps_per_second": 3.763,
+      "step": 3800
+    },
+    {
+      "epoch": 1.581356637536413,
+      "eval_international_law_loss": 3.0861918926239014,
+      "eval_international_law_runtime": 0.2673,
+      "eval_international_law_samples_per_second": 7.482,
+      "eval_international_law_steps_per_second": 3.741,
+      "step": 3800
+    },
+    {
+      "epoch": 1.5896795672076571,
+      "grad_norm": 0.34375,
+      "learning_rate": 3.9741989180191435e-06,
+      "loss": 0.2399,
+      "step": 3820
+    },
+    {
+      "epoch": 1.5980024968789013,
+      "grad_norm": 0.33984375,
+      "learning_rate": 3.995006242197254e-06,
+      "loss": 0.2298,
+      "step": 3840
+    },
+    {
+      "epoch": 1.6063254265501457,
+      "grad_norm": 0.37109375,
+      "learning_rate": 4.015813566375364e-06,
+      "loss": 0.2409,
+      "step": 3860
+    },
+    {
+      "epoch": 1.61464835622139,
+      "grad_norm": 0.36328125,
+      "learning_rate": 4.036620890553476e-06,
+      "loss": 0.2397,
+      "step": 3880
+    },
+    {
+      "epoch": 1.6229712858926342,
+      "grad_norm": 0.35546875,
+      "learning_rate": 4.057428214731586e-06,
+      "loss": 0.2373,
+      "step": 3900
+    },
+    {
+      "epoch": 1.6229712858926342,
+      "eval_main_loss": 0.2453879863023758,
+      "eval_main_runtime": 6.3211,
+      "eval_main_samples_per_second": 30.058,
+      "eval_main_steps_per_second": 3.797,
+      "step": 3900
+    },
+    {
+      "epoch": 1.6229712858926342,
+      "eval_anatomy_loss": 2.8509092330932617,
+      "eval_anatomy_runtime": 0.2658,
+      "eval_anatomy_samples_per_second": 7.525,
+      "eval_anatomy_steps_per_second": 3.762,
+      "step": 3900
+    },
+    {
+      "epoch": 1.6229712858926342,
+      "eval_college_mathematics_loss": 2.0684049129486084,
+      "eval_college_mathematics_runtime": 0.2654,
+      "eval_college_mathematics_samples_per_second": 7.535,
+      "eval_college_mathematics_steps_per_second": 3.767,
+      "step": 3900
+    },
+    {
+      "epoch": 1.6229712858926342,
+      "eval_international_law_loss": 3.0845108032226562,
+      "eval_international_law_runtime": 0.2652,
+      "eval_international_law_samples_per_second": 7.542,
+      "eval_international_law_steps_per_second": 3.771,
+      "step": 3900
+    },
+    {
+      "epoch": 1.6312942155638783,
+      "grad_norm": 0.3203125,
+      "learning_rate": 4.078235538909697e-06,
+      "loss": 0.2356,
+      "step": 3920
+    },
+    {
+      "epoch": 1.639617145235123,
+      "grad_norm": 0.36328125,
+      "learning_rate": 4.099042863087807e-06,
+      "loss": 0.238,
+      "step": 3940
+    },
+    {
+      "epoch": 1.647940074906367,
+      "grad_norm": 0.3359375,
+      "learning_rate": 4.119850187265918e-06,
+      "loss": 0.2421,
+      "step": 3960
+    },
+    {
+      "epoch": 1.6562630045776112,
+      "grad_norm": 0.380859375,
+      "learning_rate": 4.140657511444029e-06,
+      "loss": 0.2401,
+      "step": 3980
+    },
+    {
+      "epoch": 1.6645859342488556,
+      "grad_norm": 0.341796875,
+      "learning_rate": 4.161464835622139e-06,
+      "loss": 0.2372,
+      "step": 4000
+    },
+    {
+      "epoch": 1.6645859342488556,
+      "eval_main_loss": 0.2449522763490677,
+      "eval_main_runtime": 6.3294,
+      "eval_main_samples_per_second": 30.019,
+      "eval_main_steps_per_second": 3.792,
+      "step": 4000
+    },
+    {
+      "epoch": 1.6645859342488556,
+      "eval_anatomy_loss": 2.8499295711517334,
+      "eval_anatomy_runtime": 0.2671,
+      "eval_anatomy_samples_per_second": 7.487,
+      "eval_anatomy_steps_per_second": 3.744,
+      "step": 4000
+    },
+    {
+      "epoch": 1.6645859342488556,
+      "eval_college_mathematics_loss": 2.0638344287872314,
+      "eval_college_mathematics_runtime": 0.2662,
+      "eval_college_mathematics_samples_per_second": 7.514,
+      "eval_college_mathematics_steps_per_second": 3.757,
+      "step": 4000
+    },
+    {
+      "epoch": 1.6645859342488556,
+      "eval_international_law_loss": 3.08413028717041,
+      "eval_international_law_runtime": 0.2655,
+      "eval_international_law_samples_per_second": 7.534,
+      "eval_international_law_steps_per_second": 3.767,
+      "step": 4000
+    },
+    {
+      "epoch": 1.6729088639201,
+      "grad_norm": 0.361328125,
+      "learning_rate": 4.1822721598002506e-06,
+      "loss": 0.2415,
+      "step": 4020
+    },
+    {
+      "epoch": 1.6812317935913441,
+      "grad_norm": 0.349609375,
+      "learning_rate": 4.203079483978361e-06,
+      "loss": 0.244,
+      "step": 4040
+    },
+    {
+      "epoch": 1.6895547232625883,
+      "grad_norm": 0.33984375,
+      "learning_rate": 4.2238868081564715e-06,
+      "loss": 0.2371,
+      "step": 4060
+    },
+    {
+      "epoch": 1.6978776529338329,
+      "grad_norm": 0.353515625,
+      "learning_rate": 4.244694132334582e-06,
+      "loss": 0.2371,
+      "step": 4080
+    },
+    {
+      "epoch": 1.706200582605077,
+      "grad_norm": 0.318359375,
+      "learning_rate": 4.265501456512693e-06,
+      "loss": 0.2389,
+      "step": 4100
+    },
+    {
+      "epoch": 1.706200582605077,
+      "eval_main_loss": 0.24446672201156616,
+      "eval_main_runtime": 6.3267,
+      "eval_main_samples_per_second": 30.031,
+      "eval_main_steps_per_second": 3.793,
+      "step": 4100
+    },
+    {
+      "epoch": 1.706200582605077,
+      "eval_anatomy_loss": 2.8472187519073486,
+      "eval_anatomy_runtime": 0.2664,
+      "eval_anatomy_samples_per_second": 7.508,
+      "eval_anatomy_steps_per_second": 3.754,
+      "step": 4100
+    },
+    {
+      "epoch": 1.706200582605077,
+      "eval_college_mathematics_loss": 2.0650761127471924,
+      "eval_college_mathematics_runtime": 0.2662,
+      "eval_college_mathematics_samples_per_second": 7.513,
+      "eval_college_mathematics_steps_per_second": 3.757,
+      "step": 4100
+    },
+    {
+      "epoch": 1.706200582605077,
+      "eval_international_law_loss": 3.085548162460327,
+      "eval_international_law_runtime": 0.2656,
+      "eval_international_law_samples_per_second": 7.529,
+      "eval_international_law_steps_per_second": 3.765,
+      "step": 4100
+    },
+    {
+      "epoch": 1.7145235122763212,
+      "grad_norm": 0.265625,
+      "learning_rate": 4.286308780690803e-06,
+      "loss": 0.2373,
+      "step": 4120
+    },
+    {
+      "epoch": 1.7228464419475655,
+      "grad_norm": 0.330078125,
+      "learning_rate": 4.307116104868914e-06,
+      "loss": 0.2394,
+      "step": 4140
+    },
+    {
+      "epoch": 1.73116937161881,
+      "grad_norm": 0.3515625,
+      "learning_rate": 4.3279234290470254e-06,
+      "loss": 0.2376,
+      "step": 4160
+    },
+    {
+      "epoch": 1.739492301290054,
+      "grad_norm": 0.330078125,
+      "learning_rate": 4.348730753225136e-06,
+      "loss": 0.2348,
+      "step": 4180
+    },
+    {
+      "epoch": 1.7478152309612984,
+      "grad_norm": 0.392578125,
+      "learning_rate": 4.369538077403246e-06,
+      "loss": 0.2377,
+      "step": 4200
+    },
+    {
+      "epoch": 1.7478152309612984,
+      "eval_main_loss": 0.2439066469669342,
+      "eval_main_runtime": 6.3271,
+      "eval_main_samples_per_second": 30.03,
+      "eval_main_steps_per_second": 3.793,
+      "step": 4200
+    },
+    {
+      "epoch": 1.7478152309612984,
+      "eval_anatomy_loss": 2.8513143062591553,
+      "eval_anatomy_runtime": 0.2661,
+      "eval_anatomy_samples_per_second": 7.517,
+      "eval_anatomy_steps_per_second": 3.758,
+      "step": 4200
+    },
+    {
+      "epoch": 1.7478152309612984,
+      "eval_college_mathematics_loss": 2.062347412109375,
+      "eval_college_mathematics_runtime": 0.2653,
+      "eval_college_mathematics_samples_per_second": 7.539,
+      "eval_college_mathematics_steps_per_second": 3.769,
+      "step": 4200
+    },
+    {
+      "epoch": 1.7478152309612984,
+      "eval_international_law_loss": 3.082612991333008,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.504,
+      "eval_international_law_steps_per_second": 3.752,
+      "step": 4200
+    },
+    {
+      "epoch": 1.7561381606325428,
+      "grad_norm": 0.322265625,
+      "learning_rate": 4.390345401581357e-06,
+      "loss": 0.2359,
+      "step": 4220
+    },
+    {
+      "epoch": 1.764461090303787,
+      "grad_norm": 0.392578125,
+      "learning_rate": 4.411152725759468e-06,
+      "loss": 0.2387,
+      "step": 4240
+    },
+    {
+      "epoch": 1.772784019975031,
+      "grad_norm": 0.330078125,
+      "learning_rate": 4.431960049937578e-06,
+      "loss": 0.2356,
+      "step": 4260
+    },
+    {
+      "epoch": 1.7811069496462755,
+      "grad_norm": 0.326171875,
+      "learning_rate": 4.452767374115689e-06,
+      "loss": 0.236,
+      "step": 4280
+    },
+    {
+      "epoch": 1.7894298793175198,
+      "grad_norm": 0.35546875,
+      "learning_rate": 4.4735746982938e-06,
+      "loss": 0.2359,
+      "step": 4300
+    },
+    {
+      "epoch": 1.7894298793175198,
+      "eval_main_loss": 0.2437724471092224,
+      "eval_main_runtime": 6.33,
+      "eval_main_samples_per_second": 30.016,
+      "eval_main_steps_per_second": 3.791,
+      "step": 4300
+    },
+    {
+      "epoch": 1.7894298793175198,
+      "eval_anatomy_loss": 2.847775936126709,
+      "eval_anatomy_runtime": 0.2665,
+      "eval_anatomy_samples_per_second": 7.505,
+      "eval_anatomy_steps_per_second": 3.752,
+      "step": 4300
+    },
+    {
+      "epoch": 1.7894298793175198,
+      "eval_college_mathematics_loss": 2.0655105113983154,
+      "eval_college_mathematics_runtime": 0.2655,
+      "eval_college_mathematics_samples_per_second": 7.534,
+      "eval_college_mathematics_steps_per_second": 3.767,
+      "step": 4300
+    },
+    {
+      "epoch": 1.7894298793175198,
+      "eval_international_law_loss": 3.084287166595459,
+      "eval_international_law_runtime": 0.2661,
+      "eval_international_law_samples_per_second": 7.515,
+      "eval_international_law_steps_per_second": 3.757,
+      "step": 4300
+    },
+    {
+      "epoch": 1.797752808988764,
+      "grad_norm": 0.322265625,
+      "learning_rate": 4.494382022471911e-06,
+      "loss": 0.2378,
+      "step": 4320
+    },
+    {
+      "epoch": 1.8060757386600084,
+      "grad_norm": 0.322265625,
+      "learning_rate": 4.515189346650021e-06,
+      "loss": 0.2356,
+      "step": 4340
+    },
+    {
+      "epoch": 1.8143986683312527,
+      "grad_norm": 0.369140625,
+      "learning_rate": 4.535996670828132e-06,
+      "loss": 0.2358,
+      "step": 4360
+    },
+    {
+      "epoch": 1.822721598002497,
+      "grad_norm": 0.34375,
+      "learning_rate": 4.556803995006243e-06,
+      "loss": 0.2379,
+      "step": 4380
+    },
+    {
+      "epoch": 1.831044527673741,
+      "grad_norm": 0.33984375,
+      "learning_rate": 4.5776113191843534e-06,
+      "loss": 0.2395,
+      "step": 4400
+    },
+    {
+      "epoch": 1.831044527673741,
+      "eval_main_loss": 0.24319638311862946,
+      "eval_main_runtime": 6.3267,
+      "eval_main_samples_per_second": 30.032,
+      "eval_main_steps_per_second": 3.793,
+      "step": 4400
+    },
+    {
+      "epoch": 1.831044527673741,
+      "eval_anatomy_loss": 2.846129894256592,
+      "eval_anatomy_runtime": 0.2666,
+      "eval_anatomy_samples_per_second": 7.501,
+      "eval_anatomy_steps_per_second": 3.751,
+      "step": 4400
+    },
+    {
+      "epoch": 1.831044527673741,
+      "eval_college_mathematics_loss": 2.0631587505340576,
+      "eval_college_mathematics_runtime": 0.2651,
+      "eval_college_mathematics_samples_per_second": 7.545,
+      "eval_college_mathematics_steps_per_second": 3.773,
+      "step": 4400
+    },
+    {
+      "epoch": 1.831044527673741,
+      "eval_international_law_loss": 3.0830471515655518,
+      "eval_international_law_runtime": 0.2668,
+      "eval_international_law_samples_per_second": 7.498,
+      "eval_international_law_steps_per_second": 3.749,
+      "step": 4400
+    },
+    {
+      "epoch": 1.8393674573449854,
+      "grad_norm": 0.29296875,
+      "learning_rate": 4.598418643362464e-06,
+      "loss": 0.2367,
+      "step": 4420
+    },
+    {
+      "epoch": 1.8476903870162298,
+      "grad_norm": 0.31640625,
+      "learning_rate": 4.619225967540575e-06,
+      "loss": 0.2356,
+      "step": 4440
+    },
+    {
+      "epoch": 1.856013316687474,
+      "grad_norm": 0.3359375,
+      "learning_rate": 4.640033291718685e-06,
+      "loss": 0.2386,
+      "step": 4460
+    },
+    {
+      "epoch": 1.8643362463587183,
+      "grad_norm": 0.337890625,
+      "learning_rate": 4.660840615896796e-06,
+      "loss": 0.236,
+      "step": 4480
+    },
+    {
+      "epoch": 1.8726591760299627,
+      "grad_norm": 0.32421875,
+      "learning_rate": 4.6816479400749066e-06,
+      "loss": 0.24,
+      "step": 4500
+    },
+    {
+      "epoch": 1.8726591760299627,
+      "eval_main_loss": 0.2428365796804428,
+      "eval_main_runtime": 6.3298,
+      "eval_main_samples_per_second": 30.017,
+      "eval_main_steps_per_second": 3.792,
+      "step": 4500
+    },
+    {
+      "epoch": 1.8726591760299627,
+      "eval_anatomy_loss": 2.846830368041992,
+      "eval_anatomy_runtime": 0.2658,
+      "eval_anatomy_samples_per_second": 7.526,
+      "eval_anatomy_steps_per_second": 3.763,
+      "step": 4500
+    },
+    {
+      "epoch": 1.8726591760299627,
+      "eval_college_mathematics_loss": 2.061974287033081,
+      "eval_college_mathematics_runtime": 0.2663,
+      "eval_college_mathematics_samples_per_second": 7.509,
+      "eval_college_mathematics_steps_per_second": 3.755,
+      "step": 4500
+    },
+    {
+      "epoch": 1.8726591760299627,
+      "eval_international_law_loss": 3.0783193111419678,
+      "eval_international_law_runtime": 0.267,
+      "eval_international_law_samples_per_second": 7.492,
+      "eval_international_law_steps_per_second": 3.746,
+      "step": 4500
+    },
+    {
+      "epoch": 1.8809821057012068,
+      "grad_norm": 0.353515625,
+      "learning_rate": 4.702455264253018e-06,
+      "loss": 0.2359,
+      "step": 4520
+    },
+    {
+      "epoch": 1.889305035372451,
+      "grad_norm": 0.310546875,
+      "learning_rate": 4.723262588431128e-06,
+      "loss": 0.2419,
+      "step": 4540
+    },
+    {
+      "epoch": 1.8976279650436954,
+      "grad_norm": 0.294921875,
+      "learning_rate": 4.744069912609239e-06,
+      "loss": 0.2353,
+      "step": 4560
+    },
+    {
+      "epoch": 1.9059508947149397,
+      "grad_norm": 0.361328125,
+      "learning_rate": 4.76487723678735e-06,
+      "loss": 0.2364,
+      "step": 4580
+    },
+    {
+      "epoch": 1.9142738243861839,
+      "grad_norm": 0.330078125,
+      "learning_rate": 4.78568456096546e-06,
+      "loss": 0.2397,
+      "step": 4600
+    },
+    {
+      "epoch": 1.9142738243861839,
+      "eval_main_loss": 0.24256636202335358,
+      "eval_main_runtime": 6.328,
+      "eval_main_samples_per_second": 30.025,
+      "eval_main_steps_per_second": 3.793,
+      "step": 4600
+    },
+    {
+      "epoch": 1.9142738243861839,
+      "eval_anatomy_loss": 2.8468308448791504,
+      "eval_anatomy_runtime": 0.2667,
+      "eval_anatomy_samples_per_second": 7.498,
+      "eval_anatomy_steps_per_second": 3.749,
+      "step": 4600
+    },
+    {
+      "epoch": 1.9142738243861839,
+      "eval_college_mathematics_loss": 2.06019926071167,
+      "eval_college_mathematics_runtime": 0.2655,
+      "eval_college_mathematics_samples_per_second": 7.534,
+      "eval_college_mathematics_steps_per_second": 3.767,
+      "step": 4600
+    },
+    {
+      "epoch": 1.9142738243861839,
+      "eval_international_law_loss": 3.079383373260498,
+      "eval_international_law_runtime": 0.266,
+      "eval_international_law_samples_per_second": 7.518,
+      "eval_international_law_steps_per_second": 3.759,
+      "step": 4600
+    },
+    {
+      "epoch": 1.9225967540574282,
+      "grad_norm": 0.328125,
+      "learning_rate": 4.806491885143571e-06,
+      "loss": 0.2299,
+      "step": 4620
+    },
+    {
+      "epoch": 1.9309196837286726,
+      "grad_norm": 0.314453125,
+      "learning_rate": 4.8272992093216814e-06,
+      "loss": 0.2363,
+      "step": 4640
+    },
+    {
+      "epoch": 1.9392426133999168,
+      "grad_norm": 0.298828125,
+      "learning_rate": 4.848106533499793e-06,
+      "loss": 0.2336,
+      "step": 4660
+    },
+    {
+      "epoch": 1.947565543071161,
+      "grad_norm": 0.326171875,
+      "learning_rate": 4.868913857677903e-06,
+      "loss": 0.2344,
+      "step": 4680
+    },
+    {
+      "epoch": 1.9558884727424053,
+      "grad_norm": 0.296875,
+      "learning_rate": 4.889721181856014e-06,
+      "loss": 0.2373,
+      "step": 4700
+    },
+    {
+      "epoch": 1.9558884727424053,
+      "eval_main_loss": 0.2421763837337494,
+      "eval_main_runtime": 6.3346,
+      "eval_main_samples_per_second": 29.994,
+      "eval_main_steps_per_second": 3.789,
+      "step": 4700
+    },
+    {
+      "epoch": 1.9558884727424053,
+      "eval_anatomy_loss": 2.8438565731048584,
+      "eval_anatomy_runtime": 0.2673,
+      "eval_anatomy_samples_per_second": 7.483,
+      "eval_anatomy_steps_per_second": 3.742,
+      "step": 4700
+    },
+    {
+      "epoch": 1.9558884727424053,
+      "eval_college_mathematics_loss": 2.0591747760772705,
+      "eval_college_mathematics_runtime": 0.2666,
+      "eval_college_mathematics_samples_per_second": 7.501,
+      "eval_college_mathematics_steps_per_second": 3.75,
+      "step": 4700
+    },
+    {
+      "epoch": 1.9558884727424053,
+      "eval_international_law_loss": 3.0781655311584473,
+      "eval_international_law_runtime": 0.2662,
+      "eval_international_law_samples_per_second": 7.512,
+      "eval_international_law_steps_per_second": 3.756,
+      "step": 4700
+    },
+    {
+      "epoch": 1.9642114024136497,
+      "grad_norm": 0.333984375,
+      "learning_rate": 4.910528506034125e-06,
+      "loss": 0.2338,
+      "step": 4720
+    },
+    {
+      "epoch": 1.9725343320848938,
+      "grad_norm": 0.3359375,
+      "learning_rate": 4.9313358302122346e-06,
+      "loss": 0.2396,
+      "step": 4740
+    },
+    {
+      "epoch": 1.9808572617561382,
+      "grad_norm": 0.35546875,
+      "learning_rate": 4.952143154390346e-06,
+      "loss": 0.2303,
+      "step": 4760
+    },
+    {
+      "epoch": 1.9891801914273826,
+      "grad_norm": 0.310546875,
+      "learning_rate": 4.972950478568456e-06,
+      "loss": 0.2332,
+      "step": 4780
+    },
+    {
+      "epoch": 1.9975031210986267,
+      "grad_norm": 0.283203125,
+      "learning_rate": 4.993757802746567e-06,
+      "loss": 0.2409,
+      "step": 4800
+    },
+    {
+      "epoch": 1.9975031210986267,
+      "eval_main_loss": 0.24170413613319397,
+      "eval_main_runtime": 6.3319,
+      "eval_main_samples_per_second": 30.007,
+      "eval_main_steps_per_second": 3.79,
+      "step": 4800
+    },
+    {
+      "epoch": 1.9975031210986267,
+      "eval_anatomy_loss": 2.843282699584961,
+      "eval_anatomy_runtime": 0.267,
+      "eval_anatomy_samples_per_second": 7.491,
+      "eval_anatomy_steps_per_second": 3.745,
+      "step": 4800
+    },
+    {
+      "epoch": 1.9975031210986267,
+      "eval_college_mathematics_loss": 2.059091329574585,
+      "eval_college_mathematics_runtime": 0.2658,
+      "eval_college_mathematics_samples_per_second": 7.525,
+      "eval_college_mathematics_steps_per_second": 3.762,
+      "step": 4800
+    },
+    {
+      "epoch": 1.9975031210986267,
+      "eval_international_law_loss": 3.0776219367980957,
+      "eval_international_law_runtime": 0.2663,
+      "eval_international_law_samples_per_second": 7.511,
+      "eval_international_law_steps_per_second": 3.756,
+      "step": 4800
+    },
+    {
+      "epoch": 2.005826050769871,
+      "grad_norm": 0.296875,
+      "learning_rate": 4.999993456981855e-06,
+      "loss": 0.2335,
+      "step": 4820
+    },
+    {
+      "epoch": 2.0141489804411155,
+      "grad_norm": 0.3203125,
+      "learning_rate": 4.999961409628488e-06,
+      "loss": 0.2309,
+      "step": 4840
+    },
+    {
+      "epoch": 2.0224719101123596,
+      "grad_norm": 0.3515625,
+      "learning_rate": 4.999902656502973e-06,
+      "loss": 0.2327,
+      "step": 4860
+    },
+    {
+      "epoch": 2.0307948397836038,
+      "grad_norm": 0.296875,
+      "learning_rate": 4.99981719823294e-06,
+      "loss": 0.2379,
+      "step": 4880
+    },
+    {
+      "epoch": 2.039117769454848,
+      "grad_norm": 0.28515625,
+      "learning_rate": 4.999705035731294e-06,
+      "loss": 0.2372,
+      "step": 4900
+    },
+    {
+      "epoch": 2.039117769454848,
+      "eval_main_loss": 0.24145889282226562,
+      "eval_main_runtime": 6.3292,
+      "eval_main_samples_per_second": 30.02,
+      "eval_main_steps_per_second": 3.792,
+      "step": 4900
+    },
+    {
+      "epoch": 2.039117769454848,
+      "eval_anatomy_loss": 2.841543197631836,
+      "eval_anatomy_runtime": 0.2664,
+      "eval_anatomy_samples_per_second": 7.506,
+      "eval_anatomy_steps_per_second": 3.753,
+      "step": 4900
+    },
+    {
+      "epoch": 2.039117769454848,
+      "eval_college_mathematics_loss": 2.058622360229492,
+      "eval_college_mathematics_runtime": 0.2665,
+      "eval_college_mathematics_samples_per_second": 7.503,
+      "eval_college_mathematics_steps_per_second": 3.752,
+      "step": 4900
+    },
+    {
+      "epoch": 2.039117769454848,
+      "eval_international_law_loss": 3.0742015838623047,
+      "eval_international_law_runtime": 0.267,
+      "eval_international_law_samples_per_second": 7.49,
+      "eval_international_law_steps_per_second": 3.745,
+      "step": 4900
+    },
+    {
+      "epoch": 2.0474406991260925,
+      "grad_norm": 0.3203125,
+      "learning_rate": 4.999566170196208e-06,
+      "loss": 0.236,
+      "step": 4920
+    },
+    {
+      "epoch": 2.0557636287973367,
+      "grad_norm": 0.26171875,
+      "learning_rate": 4.999400603111109e-06,
+      "loss": 0.2362,
+      "step": 4940
+    },
+    {
+      "epoch": 2.064086558468581,
+      "grad_norm": 0.279296875,
+      "learning_rate": 4.999208336244664e-06,
+      "loss": 0.2338,
+      "step": 4960
+    },
+    {
+      "epoch": 2.0724094881398254,
+      "grad_norm": 0.28515625,
+      "learning_rate": 4.998989371650758e-06,
+      "loss": 0.2336,
+      "step": 4980
+    },
+    {
+      "epoch": 2.0807324178110695,
+      "grad_norm": 0.30078125,
+      "learning_rate": 4.998743711668475e-06,
+      "loss": 0.233,
+      "step": 5000
+    },
+    {
+      "epoch": 2.0807324178110695,
+      "eval_main_loss": 0.2409505397081375,
+      "eval_main_runtime": 6.3309,
+      "eval_main_samples_per_second": 30.012,
+      "eval_main_steps_per_second": 3.791,
+      "step": 5000
+    },
+    {
+      "epoch": 2.0807324178110695,
+      "eval_anatomy_loss": 2.8420751094818115,
+      "eval_anatomy_runtime": 0.2657,
+      "eval_anatomy_samples_per_second": 7.527,
+      "eval_anatomy_steps_per_second": 3.764,
+      "step": 5000
+    },
+    {
+      "epoch": 2.0807324178110695,
+      "eval_college_mathematics_loss": 2.0578598976135254,
+      "eval_college_mathematics_runtime": 0.2655,
+      "eval_college_mathematics_samples_per_second": 7.534,
+      "eval_college_mathematics_steps_per_second": 3.767,
+      "step": 5000
+    },
+    {
+      "epoch": 2.0807324178110695,
+      "eval_international_law_loss": 3.0756309032440186,
+      "eval_international_law_runtime": 0.2656,
+      "eval_international_law_samples_per_second": 7.529,
+      "eval_international_law_steps_per_second": 3.765,
+      "step": 5000
+    },
+    {
+      "epoch": 2.0890553474823137,
+      "grad_norm": 0.291015625,
+      "learning_rate": 4.998471358922071e-06,
+      "loss": 0.227,
+      "step": 5020
+    },
+    {
+      "epoch": 2.097378277153558,
+      "grad_norm": 0.275390625,
+      "learning_rate": 4.998172316320947e-06,
+      "loss": 0.2354,
+      "step": 5040
+    },
+    {
+      "epoch": 2.1057012068248024,
+      "grad_norm": 0.322265625,
+      "learning_rate": 4.997846587059618e-06,
+      "loss": 0.2314,
+      "step": 5060
+    },
+    {
+      "epoch": 2.1140241364960466,
+      "grad_norm": 0.328125,
+      "learning_rate": 4.997494174617679e-06,
+      "loss": 0.2381,
+      "step": 5080
+    },
+    {
+      "epoch": 2.1223470661672907,
+      "grad_norm": 0.28515625,
+      "learning_rate": 4.997115082759764e-06,
+      "loss": 0.233,
+      "step": 5100
+    },
+    {
+      "epoch": 2.1223470661672907,
+      "eval_main_loss": 0.2407059371471405,
+      "eval_main_runtime": 6.33,
+      "eval_main_samples_per_second": 30.016,
+      "eval_main_steps_per_second": 3.791,
+      "step": 5100
+    },
+    {
+      "epoch": 2.1223470661672907,
+      "eval_anatomy_loss": 2.8402063846588135,
+      "eval_anatomy_runtime": 0.2666,
+      "eval_anatomy_samples_per_second": 7.503,
+      "eval_anatomy_steps_per_second": 3.752,
+      "step": 5100
+    },
+    {
+      "epoch": 2.1223470661672907,
+      "eval_college_mathematics_loss": 2.0568246841430664,
+      "eval_college_mathematics_runtime": 0.2661,
+      "eval_college_mathematics_samples_per_second": 7.517,
+      "eval_college_mathematics_steps_per_second": 3.758,
+      "step": 5100
+    },
+    {
+      "epoch": 2.1223470661672907,
+      "eval_international_law_loss": 3.073444128036499,
+      "eval_international_law_runtime": 0.2662,
+      "eval_international_law_samples_per_second": 7.514,
+      "eval_international_law_steps_per_second": 3.757,
+      "step": 5100
+    },
+    {
+      "epoch": 2.1306699958385353,
+      "grad_norm": 0.322265625,
+      "learning_rate": 4.996709315535515e-06,
+      "loss": 0.2334,
+      "step": 5120
+    },
+    {
+      "epoch": 2.1389929255097795,
+      "grad_norm": 0.2578125,
+      "learning_rate": 4.9962768772795274e-06,
+      "loss": 0.235,
+      "step": 5140
+    },
+    {
+      "epoch": 2.1473158551810236,
+      "grad_norm": 0.31640625,
+      "learning_rate": 4.995817772611314e-06,
+      "loss": 0.232,
+      "step": 5160
+    },
+    {
+      "epoch": 2.1556387848522682,
+      "grad_norm": 0.3046875,
+      "learning_rate": 4.995332006435246e-06,
+      "loss": 0.2399,
+      "step": 5180
+    },
+    {
+      "epoch": 2.1639617145235124,
+      "grad_norm": 0.294921875,
+      "learning_rate": 4.9948195839405085e-06,
+      "loss": 0.2345,
+      "step": 5200
+    },
+    {
+      "epoch": 2.1639617145235124,
+      "eval_main_loss": 0.2404136210680008,
+      "eval_main_runtime": 6.3338,
+      "eval_main_samples_per_second": 29.998,
+      "eval_main_steps_per_second": 3.789,
+      "step": 5200
+    },
+    {
+      "epoch": 2.1639617145235124,
+      "eval_anatomy_loss": 2.838599443435669,
+      "eval_anatomy_runtime": 0.2657,
+      "eval_anatomy_samples_per_second": 7.527,
+      "eval_anatomy_steps_per_second": 3.763,
+      "step": 5200
+    },
+    {
+      "epoch": 2.1639617145235124,
+      "eval_college_mathematics_loss": 2.056016683578491,
+      "eval_college_mathematics_runtime": 0.2663,
+      "eval_college_mathematics_samples_per_second": 7.51,
+      "eval_college_mathematics_steps_per_second": 3.755,
+      "step": 5200
+    },
+    {
+      "epoch": 2.1639617145235124,
+      "eval_international_law_loss": 3.0730698108673096,
+      "eval_international_law_runtime": 0.2655,
+      "eval_international_law_samples_per_second": 7.533,
+      "eval_international_law_steps_per_second": 3.767,
+      "step": 5200
+    },
+    {
+      "epoch": 2.1722846441947565,
+      "grad_norm": 0.333984375,
+      "learning_rate": 4.9942805106010415e-06,
+      "loss": 0.2301,
+      "step": 5220
+    },
+    {
+      "epoch": 2.1806075738660007,
+      "grad_norm": 0.296875,
+      "learning_rate": 4.993714792175483e-06,
+      "loss": 0.2344,
+      "step": 5240
+    },
+    {
+      "epoch": 2.1889305035372453,
+      "grad_norm": 0.3125,
+      "learning_rate": 4.993122434707103e-06,
+      "loss": 0.2359,
+      "step": 5260
+    },
+    {
+      "epoch": 2.1972534332084894,
+      "grad_norm": 0.34375,
+      "learning_rate": 4.992503444523746e-06,
+      "loss": 0.2366,
+      "step": 5280
+    },
+    {
+      "epoch": 2.2055763628797336,
+      "grad_norm": 0.302734375,
+      "learning_rate": 4.991857828237757e-06,
+      "loss": 0.2369,
+      "step": 5300
+    },
+    {
+      "epoch": 2.2055763628797336,
+      "eval_main_loss": 0.24032773077487946,
+      "eval_main_runtime": 6.3258,
+      "eval_main_samples_per_second": 30.036,
+      "eval_main_steps_per_second": 3.794,
+      "step": 5300
+    },
+    {
+      "epoch": 2.2055763628797336,
+      "eval_anatomy_loss": 2.839557409286499,
+      "eval_anatomy_runtime": 0.2661,
+      "eval_anatomy_samples_per_second": 7.515,
+      "eval_anatomy_steps_per_second": 3.758,
+      "step": 5300
+    },
+    {
+      "epoch": 2.2055763628797336,
+      "eval_college_mathematics_loss": 2.0543038845062256,
+      "eval_college_mathematics_runtime": 0.2669,
+      "eval_college_mathematics_samples_per_second": 7.494,
+      "eval_college_mathematics_steps_per_second": 3.747,
+      "step": 5300
+    },
+    {
+      "epoch": 2.2055763628797336,
+      "eval_international_law_loss": 3.0714683532714844,
+      "eval_international_law_runtime": 0.2671,
+      "eval_international_law_samples_per_second": 7.487,
+      "eval_international_law_steps_per_second": 3.743,
+      "step": 5300
+    },
+    {
+      "epoch": 2.2138992925509777,
+      "grad_norm": 0.287109375,
+      "learning_rate": 4.9911855927459175e-06,
+      "loss": 0.2338,
+      "step": 5320
+    },
+    {
+      "epoch": 2.2222222222222223,
+      "grad_norm": 0.3359375,
+      "learning_rate": 4.990486745229364e-06,
+      "loss": 0.2324,
+      "step": 5340
+    },
+    {
+      "epoch": 2.2305451518934665,
+      "grad_norm": 0.337890625,
+      "learning_rate": 4.989761293153516e-06,
+      "loss": 0.2362,
+      "step": 5360
+    },
+    {
+      "epoch": 2.2388680815647106,
+      "grad_norm": 0.2734375,
+      "learning_rate": 4.989009244267998e-06,
+      "loss": 0.2331,
+      "step": 5380
+    },
+    {
+      "epoch": 2.247191011235955,
+      "grad_norm": 0.322265625,
+      "learning_rate": 4.988230606606552e-06,
+      "loss": 0.235,
+      "step": 5400
+    },
+    {
+      "epoch": 2.247191011235955,
+      "eval_main_loss": 0.2399168312549591,
+      "eval_main_runtime": 6.3585,
+      "eval_main_samples_per_second": 29.881,
+      "eval_main_steps_per_second": 3.774,
+      "step": 5400
+    },
+    {
+      "epoch": 2.247191011235955,
+      "eval_anatomy_loss": 2.83823823928833,
+      "eval_anatomy_runtime": 0.2665,
+      "eval_anatomy_samples_per_second": 7.505,
+      "eval_anatomy_steps_per_second": 3.752,
+      "step": 5400
+    },
+    {
+      "epoch": 2.247191011235955,
+      "eval_college_mathematics_loss": 2.0524001121520996,
+      "eval_college_mathematics_runtime": 0.2668,
+      "eval_college_mathematics_samples_per_second": 7.497,
+      "eval_college_mathematics_steps_per_second": 3.748,
+      "step": 5400
+    },
+    {
+      "epoch": 2.247191011235955,
+      "eval_international_law_loss": 3.073040246963501,
+      "eval_international_law_runtime": 0.2681,
+      "eval_international_law_samples_per_second": 7.461,
+      "eval_international_law_steps_per_second": 3.73,
+      "step": 5400
+    },
+    {
+      "epoch": 2.2555139409071994,
+      "grad_norm": 0.29296875,
+      "learning_rate": 4.987425388486953e-06,
+      "loss": 0.2338,
+      "step": 5420
+    },
+    {
+      "epoch": 2.2638368705784435,
+      "grad_norm": 0.29296875,
+      "learning_rate": 4.986593598510924e-06,
+      "loss": 0.2309,
+      "step": 5440
+    },
+    {
+      "epoch": 2.272159800249688,
+      "grad_norm": 0.271484375,
+      "learning_rate": 4.985735245564039e-06,
+      "loss": 0.2325,
+      "step": 5460
+    },
+    {
+      "epoch": 2.2804827299209323,
+      "grad_norm": 0.34375,
+      "learning_rate": 4.984850338815631e-06,
+      "loss": 0.2319,
+      "step": 5480
+    },
+    {
+      "epoch": 2.2888056595921764,
+      "grad_norm": 0.333984375,
+      "learning_rate": 4.983938887718692e-06,
+      "loss": 0.2414,
+      "step": 5500
+    },
+    {
+      "epoch": 2.2888056595921764,
+      "eval_main_loss": 0.23984655737876892,
+      "eval_main_runtime": 6.3511,
+      "eval_main_samples_per_second": 29.916,
+      "eval_main_steps_per_second": 3.779,
+      "step": 5500
+    },
+    {
+      "epoch": 2.2888056595921764,
+      "eval_anatomy_loss": 2.8364694118499756,
+      "eval_anatomy_runtime": 0.2673,
+      "eval_anatomy_samples_per_second": 7.483,
+      "eval_anatomy_steps_per_second": 3.741,
+      "step": 5500
+    },
+    {
+      "epoch": 2.2888056595921764,
+      "eval_college_mathematics_loss": 2.051426410675049,
+      "eval_college_mathematics_runtime": 0.2668,
+      "eval_college_mathematics_samples_per_second": 7.496,
+      "eval_college_mathematics_steps_per_second": 3.748,
+      "step": 5500
+    },
+    {
+      "epoch": 2.2888056595921764,
+      "eval_international_law_loss": 3.0698070526123047,
+      "eval_international_law_runtime": 0.2666,
+      "eval_international_law_samples_per_second": 7.502,
+      "eval_international_law_steps_per_second": 3.751,
+      "step": 5500
+    },
+    {
+      "epoch": 2.2971285892634206,
+      "grad_norm": 0.30078125,
+      "learning_rate": 4.983000902009776e-06,
+      "loss": 0.2384,
+      "step": 5520
+    },
+    {
+      "epoch": 2.305451518934665,
+      "grad_norm": 0.306640625,
+      "learning_rate": 4.982036391708891e-06,
+      "loss": 0.2342,
+      "step": 5540
+    },
+    {
+      "epoch": 2.3137744486059093,
+      "grad_norm": 0.2734375,
+      "learning_rate": 4.98104536711939e-06,
+      "loss": 0.2402,
+      "step": 5560
+    },
+    {
+      "epoch": 2.3220973782771535,
+      "grad_norm": 0.26953125,
+      "learning_rate": 4.9800278388278715e-06,
+      "loss": 0.231,
+      "step": 5580
+    },
+    {
+      "epoch": 2.3304203079483976,
+      "grad_norm": 0.294921875,
+      "learning_rate": 4.978983817704051e-06,
+      "loss": 0.2428,
+      "step": 5600
+    },
+    {
+      "epoch": 2.3304203079483976,
+      "eval_main_loss": 0.23956555128097534,
+      "eval_main_runtime": 6.334,
+      "eval_main_samples_per_second": 29.997,
+      "eval_main_steps_per_second": 3.789,
+      "step": 5600
+    },
+    {
+      "epoch": 2.3304203079483976,
+      "eval_anatomy_loss": 2.8366129398345947,
+      "eval_anatomy_runtime": 0.268,
+      "eval_anatomy_samples_per_second": 7.464,
+      "eval_anatomy_steps_per_second": 3.732,
+      "step": 5600
+    },
+    {
+      "epoch": 2.3304203079483976,
+      "eval_college_mathematics_loss": 2.0517213344573975,
+      "eval_college_mathematics_runtime": 0.2675,
+      "eval_college_mathematics_samples_per_second": 7.477,
+      "eval_college_mathematics_steps_per_second": 3.738,
+      "step": 5600
+    },
+    {
+      "epoch": 2.3304203079483976,
+      "eval_international_law_loss": 3.0719730854034424,
+      "eval_international_law_runtime": 0.2653,
+      "eval_international_law_samples_per_second": 7.538,
+      "eval_international_law_steps_per_second": 3.769,
+      "step": 5600
+    },
+    {
+      "epoch": 2.338743237619642,
+      "grad_norm": 0.259765625,
+      "learning_rate": 4.977913314900659e-06,
+      "loss": 0.228,
+      "step": 5620
+    },
+    {
+      "epoch": 2.3470661672908864,
+      "grad_norm": 0.265625,
+      "learning_rate": 4.976816341853312e-06,
+      "loss": 0.2344,
+      "step": 5640
+    },
+    {
+      "epoch": 2.3553890969621305,
+      "grad_norm": 0.294921875,
+      "learning_rate": 4.975692910280397e-06,
+      "loss": 0.2305,
+      "step": 5660
+    },
+    {
+      "epoch": 2.363712026633375,
+      "grad_norm": 0.279296875,
+      "learning_rate": 4.974543032182943e-06,
+      "loss": 0.2372,
+      "step": 5680
+    },
+    {
+      "epoch": 2.3720349563046192,
+      "grad_norm": 0.294921875,
+      "learning_rate": 4.973366719844491e-06,
+      "loss": 0.2319,
+      "step": 5700
+    },
+    {
+      "epoch": 2.3720349563046192,
+      "eval_main_loss": 0.23942260444164276,
+      "eval_main_runtime": 6.3369,
+      "eval_main_samples_per_second": 29.983,
+      "eval_main_steps_per_second": 3.787,
+      "step": 5700
+    },
+    {
+      "epoch": 2.3720349563046192,
+      "eval_anatomy_loss": 2.8341376781463623,
+      "eval_anatomy_runtime": 0.2654,
+      "eval_anatomy_samples_per_second": 7.537,
+      "eval_anatomy_steps_per_second": 3.768,
+      "step": 5700
+    },
+    {
+      "epoch": 2.3720349563046192,
+      "eval_college_mathematics_loss": 2.0530576705932617,
+      "eval_college_mathematics_runtime": 0.2663,
+      "eval_college_mathematics_samples_per_second": 7.512,
+      "eval_college_mathematics_steps_per_second": 3.756,
+      "step": 5700
+    },
+    {
+      "epoch": 2.3720349563046192,
+      "eval_international_law_loss": 3.070324420928955,
+      "eval_international_law_runtime": 0.2662,
+      "eval_international_law_samples_per_second": 7.513,
+      "eval_international_law_steps_per_second": 3.757,
+      "step": 5700
+    },
+    {
+      "epoch": 2.3803578859758634,
+      "grad_norm": 0.30859375,
+      "learning_rate": 4.972163985830967e-06,
+      "loss": 0.2326,
+      "step": 5720
+    },
+    {
+      "epoch": 2.388680815647108,
+      "grad_norm": 0.296875,
+      "learning_rate": 4.970934842990546e-06,
+      "loss": 0.234,
+      "step": 5740
+    },
+    {
+      "epoch": 2.397003745318352,
+      "grad_norm": 0.28515625,
+      "learning_rate": 4.969679304453513e-06,
+      "loss": 0.231,
+      "step": 5760
+    },
+    {
+      "epoch": 2.4053266749895963,
+      "grad_norm": 0.33984375,
+      "learning_rate": 4.968397383632127e-06,
+      "loss": 0.2322,
+      "step": 5780
+    },
+    {
+      "epoch": 2.4136496046608404,
+      "grad_norm": 0.263671875,
+      "learning_rate": 4.967089094220473e-06,
+      "loss": 0.2276,
+      "step": 5800
+    },
+    {
+      "epoch": 2.4136496046608404,
+      "eval_main_loss": 0.23935824632644653,
+      "eval_main_runtime": 6.3314,
+      "eval_main_samples_per_second": 30.009,
+      "eval_main_steps_per_second": 3.791,
+      "step": 5800
+    },
+    {
+      "epoch": 2.4136496046608404,
+      "eval_anatomy_loss": 2.8363730907440186,
+      "eval_anatomy_runtime": 0.2672,
+      "eval_anatomy_samples_per_second": 7.484,
+      "eval_anatomy_steps_per_second": 3.742,
+      "step": 5800
+    },
+    {
+      "epoch": 2.4136496046608404,
+      "eval_college_mathematics_loss": 2.04972767829895,
+      "eval_college_mathematics_runtime": 0.2664,
+      "eval_college_mathematics_samples_per_second": 7.506,
+      "eval_college_mathematics_steps_per_second": 3.753,
+      "step": 5800
+    },
+    {
+      "epoch": 2.4136496046608404,
+      "eval_international_law_loss": 3.0698299407958984,
+      "eval_international_law_runtime": 0.2654,
+      "eval_international_law_samples_per_second": 7.536,
+      "eval_international_law_steps_per_second": 3.768,
+      "step": 5800
+    },
+    {
+      "epoch": 2.421972534332085,
+      "grad_norm": 0.2734375,
+      "learning_rate": 4.9657544501943175e-06,
+      "loss": 0.2349,
+      "step": 5820
+    },
+    {
+      "epoch": 2.430295464003329,
+      "grad_norm": 0.28125,
+      "learning_rate": 4.964393465810963e-06,
+      "loss": 0.236,
+      "step": 5840
+    },
+    {
+      "epoch": 2.4386183936745733,
+      "grad_norm": 0.248046875,
+      "learning_rate": 4.9630061556090855e-06,
+      "loss": 0.2362,
+      "step": 5860
+    },
+    {
+      "epoch": 2.4469413233458175,
+      "grad_norm": 0.240234375,
+      "learning_rate": 4.961592534408592e-06,
+      "loss": 0.2285,
+      "step": 5880
+    },
+    {
+      "epoch": 2.455264253017062,
+      "grad_norm": 0.30859375,
+      "learning_rate": 4.9601526173104544e-06,
+      "loss": 0.2346,
+      "step": 5900
+    },
+    {
+      "epoch": 2.455264253017062,
+      "eval_main_loss": 0.2392611801624298,
+      "eval_main_runtime": 6.3213,
+      "eval_main_samples_per_second": 30.057,
+      "eval_main_steps_per_second": 3.797,
+      "step": 5900
+    },
+    {
+      "epoch": 2.455264253017062,
+      "eval_anatomy_loss": 2.8343961238861084,
+      "eval_anatomy_runtime": 0.2665,
+      "eval_anatomy_samples_per_second": 7.505,
+      "eval_anatomy_steps_per_second": 3.753,
+      "step": 5900
+    },
+    {
+      "epoch": 2.455264253017062,
+      "eval_college_mathematics_loss": 2.0529263019561768,
+      "eval_college_mathematics_runtime": 0.2662,
+      "eval_college_mathematics_samples_per_second": 7.514,
+      "eval_college_mathematics_steps_per_second": 3.757,
+      "step": 5900
+    },
+    {
+      "epoch": 2.455264253017062,
+      "eval_international_law_loss": 3.068134069442749,
+      "eval_international_law_runtime": 0.2664,
+      "eval_international_law_samples_per_second": 7.507,
+      "eval_international_law_steps_per_second": 3.753,
+      "step": 5900
+    },
+    {
+      "epoch": 2.4635871826883062,
+      "grad_norm": 0.291015625,
+      "learning_rate": 4.958686419696548e-06,
+      "loss": 0.228,
+      "step": 5920
+    },
+    {
+      "epoch": 2.4719101123595504,
+      "grad_norm": 0.265625,
+      "learning_rate": 4.9571939572294914e-06,
+      "loss": 0.233,
+      "step": 5940
+    },
+    {
+      "epoch": 2.480233042030795,
+      "grad_norm": 0.259765625,
+      "learning_rate": 4.955675245852475e-06,
+      "loss": 0.2316,
+      "step": 5960
+    },
+    {
+      "epoch": 2.488555971702039,
+      "grad_norm": 0.33984375,
+      "learning_rate": 4.954130301789093e-06,
+      "loss": 0.2342,
+      "step": 5980
+    },
+    {
+      "epoch": 2.4968789013732833,
+      "grad_norm": 0.28125,
+      "learning_rate": 4.952559141543171e-06,
+      "loss": 0.233,
+      "step": 6000
+    },
+    {
+      "epoch": 2.4968789013732833,
+      "eval_main_loss": 0.23919633030891418,
+      "eval_main_runtime": 6.3217,
+      "eval_main_samples_per_second": 30.055,
+      "eval_main_steps_per_second": 3.796,
+      "step": 6000
+    },
+    {
+      "epoch": 2.4968789013732833,
+      "eval_anatomy_loss": 2.835357427597046,
+      "eval_anatomy_runtime": 0.2661,
+      "eval_anatomy_samples_per_second": 7.515,
+      "eval_anatomy_steps_per_second": 3.758,
+      "step": 6000
+    },
+    {
+      "epoch": 2.4968789013732833,
+      "eval_college_mathematics_loss": 2.0497984886169434,
+      "eval_college_mathematics_runtime": 0.2664,
+      "eval_college_mathematics_samples_per_second": 7.508,
+      "eval_college_mathematics_steps_per_second": 3.754,
+      "step": 6000
+    },
+    {
+      "epoch": 2.4968789013732833,
+      "eval_international_law_loss": 3.0689234733581543,
+      "eval_international_law_runtime": 0.2669,
+      "eval_international_law_samples_per_second": 7.494,
+      "eval_international_law_steps_per_second": 3.747,
+      "step": 6000
+    },
+    {
+      "epoch": 2.505201831044528,
+      "grad_norm": 0.271484375,
+      "learning_rate": 4.950961781898586e-06,
+      "loss": 0.2345,
+      "step": 6020
+    },
+    {
+      "epoch": 2.513524760715772,
+      "grad_norm": 0.2578125,
+      "learning_rate": 4.94933823991909e-06,
+      "loss": 0.2327,
+      "step": 6040
+    },
+    {
+      "epoch": 2.521847690387016,
+      "grad_norm": 0.34765625,
+      "learning_rate": 4.947688532948129e-06,
+      "loss": 0.2286,
+      "step": 6060
+    },
+    {
+      "epoch": 2.5301706200582608,
+      "grad_norm": 0.283203125,
+      "learning_rate": 4.9460126786086535e-06,
+      "loss": 0.2313,
+      "step": 6080
+    },
+    {
+      "epoch": 2.538493549729505,
+      "grad_norm": 0.25,
+      "learning_rate": 4.944310694802935e-06,
+      "loss": 0.2343,
+      "step": 6100
+    },
+    {
+      "epoch": 2.538493549729505,
+      "eval_main_loss": 0.23911188542842865,
+      "eval_main_runtime": 6.33,
+      "eval_main_samples_per_second": 30.016,
+      "eval_main_steps_per_second": 3.791,
+      "step": 6100
+    },
+    {
+      "epoch": 2.538493549729505,
+      "eval_anatomy_loss": 2.834503650665283,
+      "eval_anatomy_runtime": 0.2651,
+      "eval_anatomy_samples_per_second": 7.545,
+      "eval_anatomy_steps_per_second": 3.772,
+      "step": 6100
+    },
+    {
+      "epoch": 2.538493549729505,
+      "eval_college_mathematics_loss": 2.049175977706909,
+      "eval_college_mathematics_runtime": 0.2667,
+      "eval_college_mathematics_samples_per_second": 7.5,
+      "eval_college_mathematics_steps_per_second": 3.75,
+      "step": 6100
+    },
+    {
+      "epoch": 2.538493549729505,
+      "eval_international_law_loss": 3.0691072940826416,
+      "eval_international_law_runtime": 0.2666,
+      "eval_international_law_samples_per_second": 7.502,
+      "eval_international_law_steps_per_second": 3.751,
+      "step": 6100
+    },
+    {
+      "epoch": 2.546816479400749,
+      "grad_norm": 0.29296875,
+      "learning_rate": 4.942582599712369e-06,
+      "loss": 0.2356,
+      "step": 6120
+    },
+    {
+      "epoch": 2.555139409071993,
+      "grad_norm": 0.29296875,
+      "learning_rate": 4.940828411797287e-06,
+      "loss": 0.2369,
+      "step": 6140
+    },
+    {
+      "epoch": 2.5634623387432374,
+      "grad_norm": 0.306640625,
+      "learning_rate": 4.9390481497967545e-06,
+      "loss": 0.2287,
+      "step": 6160
+    },
+    {
+      "epoch": 2.571785268414482,
+      "grad_norm": 0.298828125,
+      "learning_rate": 4.937241832728373e-06,
+      "loss": 0.2288,
+      "step": 6180
+    },
+    {
+      "epoch": 2.580108198085726,
+      "grad_norm": 0.2890625,
+      "learning_rate": 4.9354094798880806e-06,
+      "loss": 0.2301,
+      "step": 6200
+    },
+    {
+      "epoch": 2.580108198085726,
+      "eval_main_loss": 0.23901519179344177,
+      "eval_main_runtime": 6.3573,
+      "eval_main_samples_per_second": 29.887,
+      "eval_main_steps_per_second": 3.775,
+      "step": 6200
+    },
+    {
+      "epoch": 2.580108198085726,
+      "eval_anatomy_loss": 2.8348350524902344,
+      "eval_anatomy_runtime": 0.2662,
+      "eval_anatomy_samples_per_second": 7.513,
+      "eval_anatomy_steps_per_second": 3.757,
+      "step": 6200
+    },
+    {
+      "epoch": 2.580108198085726,
+      "eval_college_mathematics_loss": 2.048042058944702,
+      "eval_college_mathematics_runtime": 0.2678,
+      "eval_college_mathematics_samples_per_second": 7.467,
+      "eval_college_mathematics_steps_per_second": 3.734,
+      "step": 6200
+    },
+    {
+      "epoch": 2.580108198085726,
+      "eval_international_law_loss": 3.0679852962493896,
+      "eval_international_law_runtime": 0.2673,
+      "eval_international_law_samples_per_second": 7.483,
+      "eval_international_law_steps_per_second": 3.742,
+      "step": 6200
+    },
+    {
+      "epoch": 2.5884311277569703,
+      "grad_norm": 0.251953125,
+      "learning_rate": 4.9335511108499344e-06,
+      "loss": 0.2333,
+      "step": 6220
+    },
+    {
+      "epoch": 2.596754057428215,
+      "grad_norm": 0.291015625,
+      "learning_rate": 4.931666745465915e-06,
+      "loss": 0.2275,
+      "step": 6240
+    },
+    {
+      "epoch": 2.605076987099459,
+      "grad_norm": 0.26953125,
+      "learning_rate": 4.929756403865706e-06,
+      "loss": 0.2303,
+      "step": 6260
+    },
+    {
+      "epoch": 2.613399916770703,
+      "grad_norm": 0.251953125,
+      "learning_rate": 4.927820106456481e-06,
+      "loss": 0.2314,
+      "step": 6280
+    },
+    {
+      "epoch": 2.6217228464419478,
+      "grad_norm": 0.2734375,
+      "learning_rate": 4.925857873922686e-06,
+      "loss": 0.2344,
+      "step": 6300
+    },
+    {
+      "epoch": 2.6217228464419478,
+      "eval_main_loss": 0.2388763278722763,
+      "eval_main_runtime": 6.3519,
+      "eval_main_samples_per_second": 29.912,
+      "eval_main_steps_per_second": 3.778,
+      "step": 6300
+    },
+    {
+      "epoch": 2.6217228464419478,
+      "eval_anatomy_loss": 2.8354506492614746,
+      "eval_anatomy_runtime": 0.2675,
+      "eval_anatomy_samples_per_second": 7.478,
+      "eval_anatomy_steps_per_second": 3.739,
+      "step": 6300
+    },
+    {
+      "epoch": 2.6217228464419478,
+      "eval_college_mathematics_loss": 2.0530946254730225,
+      "eval_college_mathematics_runtime": 0.2664,
+      "eval_college_mathematics_samples_per_second": 7.508,
+      "eval_college_mathematics_steps_per_second": 3.754,
+      "step": 6300
+    },
+    {
+      "epoch": 2.6217228464419478,
+      "eval_international_law_loss": 3.0681755542755127,
+      "eval_international_law_runtime": 0.2672,
+      "eval_international_law_samples_per_second": 7.485,
+      "eval_international_law_steps_per_second": 3.742,
+      "step": 6300
+    },
+    {
+      "epoch": 2.630045776113192,
+      "grad_norm": 0.287109375,
+      "learning_rate": 4.923869727225819e-06,
+      "loss": 0.23,
+      "step": 6320
+    },
+    {
+      "epoch": 2.638368705784436,
+      "grad_norm": 0.259765625,
+      "learning_rate": 4.921855687604206e-06,
+      "loss": 0.2337,
+      "step": 6340
+    },
+    {
+      "epoch": 2.6466916354556806,
+      "grad_norm": 0.267578125,
+      "learning_rate": 4.91981577657277e-06,
+      "loss": 0.2337,
+      "step": 6360
+    },
+    {
+      "epoch": 2.655014565126925,
+      "grad_norm": 0.2412109375,
+      "learning_rate": 4.917750015922809e-06,
+      "loss": 0.2363,
+      "step": 6380
+    },
+    {
+      "epoch": 2.663337494798169,
+      "grad_norm": 0.29296875,
+      "learning_rate": 4.915658427721755e-06,
+      "loss": 0.2317,
+      "step": 6400
+    },
+    {
+      "epoch": 2.663337494798169,
+      "eval_main_loss": 0.23897981643676758,
+      "eval_main_runtime": 6.3411,
+      "eval_main_samples_per_second": 29.963,
+      "eval_main_steps_per_second": 3.785,
+      "step": 6400
+    },
+    {
+      "epoch": 2.663337494798169,
+      "eval_anatomy_loss": 2.8342771530151367,
+      "eval_anatomy_runtime": 0.2668,
+      "eval_anatomy_samples_per_second": 7.495,
+      "eval_anatomy_steps_per_second": 3.748,
+      "step": 6400
+    },
+    {
+      "epoch": 2.663337494798169,
+      "eval_college_mathematics_loss": 2.0512688159942627,
+      "eval_college_mathematics_runtime": 0.2663,
+      "eval_college_mathematics_samples_per_second": 7.51,
+      "eval_college_mathematics_steps_per_second": 3.755,
+      "step": 6400
+    },
+    {
+      "epoch": 2.663337494798169,
+      "eval_international_law_loss": 3.0684025287628174,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.504,
+      "eval_international_law_steps_per_second": 3.752,
+      "step": 6400
+    },
+    {
+      "epoch": 2.671660424469413,
+      "grad_norm": 0.310546875,
+      "learning_rate": 4.9135410343129465e-06,
+      "loss": 0.2327,
+      "step": 6420
+    },
+    {
+      "epoch": 2.6799833541406572,
+      "grad_norm": 0.306640625,
+      "learning_rate": 4.911397858315382e-06,
+      "loss": 0.2279,
+      "step": 6440
+    },
+    {
+      "epoch": 2.688306283811902,
+      "grad_norm": 0.326171875,
+      "learning_rate": 4.909228922623482e-06,
+      "loss": 0.2287,
+      "step": 6460
+    },
+    {
+      "epoch": 2.696629213483146,
+      "grad_norm": 0.337890625,
+      "learning_rate": 4.907034250406846e-06,
+      "loss": 0.2288,
+      "step": 6480
+    },
+    {
+      "epoch": 2.70495214315439,
+      "grad_norm": 0.294921875,
+      "learning_rate": 4.904813865110002e-06,
+      "loss": 0.2308,
+      "step": 6500
+    },
+    {
+      "epoch": 2.70495214315439,
+      "eval_main_loss": 0.23884011805057526,
+      "eval_main_runtime": 6.3205,
+      "eval_main_samples_per_second": 30.061,
+      "eval_main_steps_per_second": 3.797,
+      "step": 6500
+    },
+    {
+      "epoch": 2.70495214315439,
+      "eval_anatomy_loss": 2.8351662158966064,
+      "eval_anatomy_runtime": 0.2663,
+      "eval_anatomy_samples_per_second": 7.51,
+      "eval_anatomy_steps_per_second": 3.755,
+      "step": 6500
+    },
+    {
+      "epoch": 2.70495214315439,
+      "eval_college_mathematics_loss": 2.0514416694641113,
+      "eval_college_mathematics_runtime": 0.2668,
+      "eval_college_mathematics_samples_per_second": 7.496,
+      "eval_college_mathematics_steps_per_second": 3.748,
+      "step": 6500
+    },
+    {
+      "epoch": 2.70495214315439,
+      "eval_international_law_loss": 3.0670132637023926,
+      "eval_international_law_runtime": 0.265,
+      "eval_international_law_samples_per_second": 7.546,
+      "eval_international_law_steps_per_second": 3.773,
+      "step": 6500
+    },
+    {
+      "epoch": 2.7132750728256347,
+      "grad_norm": 0.302734375,
+      "learning_rate": 4.902567790452158e-06,
+      "loss": 0.2308,
+      "step": 6520
+    },
+    {
+      "epoch": 2.721598002496879,
+      "grad_norm": 0.263671875,
+      "learning_rate": 4.900296050426947e-06,
+      "loss": 0.2373,
+      "step": 6540
+    },
+    {
+      "epoch": 2.729920932168123,
+      "grad_norm": 0.2197265625,
+      "learning_rate": 4.897998669302173e-06,
+      "loss": 0.2291,
+      "step": 6560
+    },
+    {
+      "epoch": 2.7382438618393676,
+      "grad_norm": 0.36328125,
+      "learning_rate": 4.895675671619549e-06,
+      "loss": 0.2369,
+      "step": 6580
+    },
+    {
+      "epoch": 2.746566791510612,
+      "grad_norm": 0.2578125,
+      "learning_rate": 4.893327082194436e-06,
+      "loss": 0.2353,
+      "step": 6600
+    },
+    {
+      "epoch": 2.746566791510612,
+      "eval_main_loss": 0.23877017199993134,
+      "eval_main_runtime": 6.35,
+      "eval_main_samples_per_second": 29.921,
+      "eval_main_steps_per_second": 3.78,
+      "step": 6600
+    },
+    {
+      "epoch": 2.746566791510612,
+      "eval_anatomy_loss": 2.832685947418213,
+      "eval_anatomy_runtime": 0.2666,
+      "eval_anatomy_samples_per_second": 7.502,
+      "eval_anatomy_steps_per_second": 3.751,
+      "step": 6600
+    },
+    {
+      "epoch": 2.746566791510612,
+      "eval_college_mathematics_loss": 2.0484695434570312,
+      "eval_college_mathematics_runtime": 0.2671,
+      "eval_college_mathematics_samples_per_second": 7.489,
+      "eval_college_mathematics_steps_per_second": 3.744,
+      "step": 6600
+    },
+    {
+      "epoch": 2.746566791510612,
+      "eval_international_law_loss": 3.0677947998046875,
+      "eval_international_law_runtime": 0.2669,
+      "eval_international_law_samples_per_second": 7.493,
+      "eval_international_law_steps_per_second": 3.746,
+      "step": 6600
+    },
+    {
+      "epoch": 2.754889721181856,
+      "grad_norm": 0.267578125,
+      "learning_rate": 4.890952926115581e-06,
+      "loss": 0.2308,
+      "step": 6620
+    },
+    {
+      "epoch": 2.7632126508531005,
+      "grad_norm": 0.2890625,
+      "learning_rate": 4.888553228744842e-06,
+      "loss": 0.233,
+      "step": 6640
+    },
+    {
+      "epoch": 2.7715355805243447,
+      "grad_norm": 0.306640625,
+      "learning_rate": 4.886128015716925e-06,
+      "loss": 0.2397,
+      "step": 6660
+    },
+    {
+      "epoch": 2.779858510195589,
+      "grad_norm": 0.3203125,
+      "learning_rate": 4.883677312939103e-06,
+      "loss": 0.2345,
+      "step": 6680
+    },
+    {
+      "epoch": 2.788181439866833,
+      "grad_norm": 0.291015625,
+      "learning_rate": 4.881201146590945e-06,
+      "loss": 0.2314,
+      "step": 6700
+    },
+    {
+      "epoch": 2.788181439866833,
+      "eval_main_loss": 0.23880185186862946,
+      "eval_main_runtime": 6.3571,
+      "eval_main_samples_per_second": 29.888,
+      "eval_main_steps_per_second": 3.775,
+      "step": 6700
+    },
+    {
+      "epoch": 2.788181439866833,
+      "eval_anatomy_loss": 2.83292555809021,
+      "eval_anatomy_runtime": 0.2673,
+      "eval_anatomy_samples_per_second": 7.483,
+      "eval_anatomy_steps_per_second": 3.741,
+      "step": 6700
+    },
+    {
+      "epoch": 2.788181439866833,
+      "eval_college_mathematics_loss": 2.048858880996704,
+      "eval_college_mathematics_runtime": 0.2675,
+      "eval_college_mathematics_samples_per_second": 7.478,
+      "eval_college_mathematics_steps_per_second": 3.739,
+      "step": 6700
+    },
+    {
+      "epoch": 2.788181439866833,
+      "eval_international_law_loss": 3.066660165786743,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.505,
+      "eval_international_law_steps_per_second": 3.752,
+      "step": 6700
+    },
+    {
+      "epoch": 2.796504369538077,
+      "grad_norm": 0.251953125,
+      "learning_rate": 4.878699543124031e-06,
+      "loss": 0.2328,
+      "step": 6720
+    },
+    {
+      "epoch": 2.8048272992093217,
+      "grad_norm": 0.291015625,
+      "learning_rate": 4.876172529261678e-06,
+      "loss": 0.2334,
+      "step": 6740
+    },
+    {
+      "epoch": 2.813150228880566,
+      "grad_norm": 0.28515625,
+      "learning_rate": 4.873620131998642e-06,
+      "loss": 0.2327,
+      "step": 6760
+    },
+    {
+      "epoch": 2.82147315855181,
+      "grad_norm": 0.314453125,
+      "learning_rate": 4.871042378600842e-06,
+      "loss": 0.2309,
+      "step": 6780
+    },
+    {
+      "epoch": 2.8297960882230546,
+      "grad_norm": 0.2890625,
+      "learning_rate": 4.8684392966050594e-06,
+      "loss": 0.2307,
+      "step": 6800
+    },
+    {
+      "epoch": 2.8297960882230546,
+      "eval_main_loss": 0.2386612743139267,
+      "eval_main_runtime": 6.3547,
+      "eval_main_samples_per_second": 29.899,
+      "eval_main_steps_per_second": 3.777,
+      "step": 6800
+    },
+    {
+      "epoch": 2.8297960882230546,
+      "eval_anatomy_loss": 2.83203125,
+      "eval_anatomy_runtime": 0.268,
+      "eval_anatomy_samples_per_second": 7.464,
+      "eval_anatomy_steps_per_second": 3.732,
+      "step": 6800
+    },
+    {
+      "epoch": 2.8297960882230546,
+      "eval_college_mathematics_loss": 2.0493810176849365,
+      "eval_college_mathematics_runtime": 0.2671,
+      "eval_college_mathematics_samples_per_second": 7.487,
+      "eval_college_mathematics_steps_per_second": 3.743,
+      "step": 6800
+    },
+    {
+      "epoch": 2.8297960882230546,
+      "eval_international_law_loss": 3.065821409225464,
+      "eval_international_law_runtime": 0.2667,
+      "eval_international_law_samples_per_second": 7.498,
+      "eval_international_law_steps_per_second": 3.749,
+      "step": 6800
+    },
+    {
+      "epoch": 2.8381190178942988,
+      "grad_norm": 0.302734375,
+      "learning_rate": 4.865810913818651e-06,
+      "loss": 0.2294,
+      "step": 6820
+    },
+    {
+      "epoch": 2.846441947565543,
+      "grad_norm": 0.291015625,
+      "learning_rate": 4.863157258319245e-06,
+      "loss": 0.2324,
+      "step": 6840
+    },
+    {
+      "epoch": 2.8547648772367875,
+      "grad_norm": 0.27734375,
+      "learning_rate": 4.8604783584544475e-06,
+      "loss": 0.2313,
+      "step": 6860
+    },
+    {
+      "epoch": 2.8630878069080317,
+      "grad_norm": 0.28515625,
+      "learning_rate": 4.857774242841536e-06,
+      "loss": 0.2281,
+      "step": 6880
+    },
+    {
+      "epoch": 2.871410736579276,
+      "grad_norm": 0.296875,
+      "learning_rate": 4.855044940367155e-06,
+      "loss": 0.2329,
+      "step": 6900
+    },
+    {
+      "epoch": 2.871410736579276,
+      "eval_main_loss": 0.23859336972236633,
+      "eval_main_runtime": 6.3286,
+      "eval_main_samples_per_second": 30.022,
+      "eval_main_steps_per_second": 3.792,
+      "step": 6900
+    },
+    {
+      "epoch": 2.871410736579276,
+      "eval_anatomy_loss": 2.833575487136841,
+      "eval_anatomy_runtime": 0.2664,
+      "eval_anatomy_samples_per_second": 7.507,
+      "eval_anatomy_steps_per_second": 3.753,
+      "step": 6900
+    },
+    {
+      "epoch": 2.871410736579276,
+      "eval_college_mathematics_loss": 2.0479042530059814,
+      "eval_college_mathematics_runtime": 0.2652,
+      "eval_college_mathematics_samples_per_second": 7.542,
+      "eval_college_mathematics_steps_per_second": 3.771,
+      "step": 6900
+    },
+    {
+      "epoch": 2.871410736579276,
+      "eval_international_law_loss": 3.0679168701171875,
+      "eval_international_law_runtime": 0.2661,
+      "eval_international_law_samples_per_second": 7.516,
+      "eval_international_law_steps_per_second": 3.758,
+      "step": 6900
+    },
+    {
+      "epoch": 2.8797336662505204,
+      "grad_norm": 0.275390625,
+      "learning_rate": 4.8522904801870065e-06,
+      "loss": 0.2309,
+      "step": 6920
+    },
+    {
+      "epoch": 2.8880565959217646,
+      "grad_norm": 0.330078125,
+      "learning_rate": 4.8495108917255385e-06,
+      "loss": 0.2347,
+      "step": 6940
+    },
+    {
+      "epoch": 2.8963795255930087,
+      "grad_norm": 0.296875,
+      "learning_rate": 4.846706204675632e-06,
+      "loss": 0.2301,
+      "step": 6960
+    },
+    {
+      "epoch": 2.904702455264253,
+      "grad_norm": 0.35546875,
+      "learning_rate": 4.843876448998283e-06,
+      "loss": 0.2354,
+      "step": 6980
+    },
+    {
+      "epoch": 2.9130253849354975,
+      "grad_norm": 0.314453125,
+      "learning_rate": 4.841021654922281e-06,
+      "loss": 0.2359,
+      "step": 7000
+    },
+    {
+      "epoch": 2.9130253849354975,
+      "eval_main_loss": 0.23874352872371674,
+      "eval_main_runtime": 6.316,
+      "eval_main_samples_per_second": 30.083,
+      "eval_main_steps_per_second": 3.8,
+      "step": 7000
+    },
+    {
+      "epoch": 2.9130253849354975,
+      "eval_anatomy_loss": 2.832608699798584,
+      "eval_anatomy_runtime": 0.2648,
+      "eval_anatomy_samples_per_second": 7.552,
+      "eval_anatomy_steps_per_second": 3.776,
+      "step": 7000
+    },
+    {
+      "epoch": 2.9130253849354975,
+      "eval_college_mathematics_loss": 2.0510120391845703,
+      "eval_college_mathematics_runtime": 0.2649,
+      "eval_college_mathematics_samples_per_second": 7.55,
+      "eval_college_mathematics_steps_per_second": 3.775,
+      "step": 7000
+    },
+    {
+      "epoch": 2.9130253849354975,
+      "eval_international_law_loss": 3.0677924156188965,
+      "eval_international_law_runtime": 0.2642,
+      "eval_international_law_samples_per_second": 7.57,
+      "eval_international_law_steps_per_second": 3.785,
+      "step": 7000
+    },
+    {
+      "epoch": 2.9213483146067416,
+      "grad_norm": 0.298828125,
+      "learning_rate": 4.838141852943891e-06,
+      "loss": 0.2289,
+      "step": 7020
+    },
+    {
+      "epoch": 2.9296712442779858,
+      "grad_norm": 0.330078125,
+      "learning_rate": 4.835237073826521e-06,
+      "loss": 0.2305,
+      "step": 7040
+    },
+    {
+      "epoch": 2.93799417394923,
+      "grad_norm": 0.279296875,
+      "learning_rate": 4.8323073486003976e-06,
+      "loss": 0.2312,
+      "step": 7060
+    },
+    {
+      "epoch": 2.9463171036204745,
+      "grad_norm": 0.28515625,
+      "learning_rate": 4.829352708562233e-06,
+      "loss": 0.2321,
+      "step": 7080
+    },
+    {
+      "epoch": 2.9546400332917186,
+      "grad_norm": 0.296875,
+      "learning_rate": 4.826373185274893e-06,
+      "loss": 0.2336,
+      "step": 7100
+    },
+    {
+      "epoch": 2.9546400332917186,
+      "eval_main_loss": 0.23865433037281036,
+      "eval_main_runtime": 6.3269,
+      "eval_main_samples_per_second": 30.031,
+      "eval_main_steps_per_second": 3.793,
+      "step": 7100
+    },
+    {
+      "epoch": 2.9546400332917186,
+      "eval_anatomy_loss": 2.8344128131866455,
+      "eval_anatomy_runtime": 0.2662,
+      "eval_anatomy_samples_per_second": 7.513,
+      "eval_anatomy_steps_per_second": 3.756,
+      "step": 7100
+    },
+    {
+      "epoch": 2.9546400332917186,
+      "eval_college_mathematics_loss": 2.0521576404571533,
+      "eval_college_mathematics_runtime": 0.2649,
+      "eval_college_mathematics_samples_per_second": 7.55,
+      "eval_college_mathematics_steps_per_second": 3.775,
+      "step": 7100
+    },
+    {
+      "epoch": 2.9546400332917186,
+      "eval_international_law_loss": 3.0670950412750244,
+      "eval_international_law_runtime": 0.2651,
+      "eval_international_law_samples_per_second": 7.545,
+      "eval_international_law_steps_per_second": 3.772,
+      "step": 7100
+    },
+    {
+      "epoch": 2.962962962962963,
+      "grad_norm": 0.267578125,
+      "learning_rate": 4.823368810567056e-06,
+      "loss": 0.2332,
+      "step": 7120
+    },
+    {
+      "epoch": 2.9712858926342074,
+      "grad_norm": 0.31640625,
+      "learning_rate": 4.820339616532878e-06,
+      "loss": 0.2297,
+      "step": 7140
+    },
+    {
+      "epoch": 2.9796088223054515,
+      "grad_norm": 0.283203125,
+      "learning_rate": 4.817285635531641e-06,
+      "loss": 0.2311,
+      "step": 7160
+    },
+    {
+      "epoch": 2.9879317519766957,
+      "grad_norm": 0.30859375,
+      "learning_rate": 4.81420690018742e-06,
+      "loss": 0.231,
+      "step": 7180
+    },
+    {
+      "epoch": 2.9962546816479403,
+      "grad_norm": 0.31640625,
+      "learning_rate": 4.811103443388724e-06,
+      "loss": 0.2332,
+      "step": 7200
+    },
+    {
+      "epoch": 2.9962546816479403,
+      "eval_main_loss": 0.23862618207931519,
+      "eval_main_runtime": 6.3548,
+      "eval_main_samples_per_second": 29.899,
+      "eval_main_steps_per_second": 3.777,
+      "step": 7200
+    },
+    {
+      "epoch": 2.9962546816479403,
+      "eval_anatomy_loss": 2.834916114807129,
+      "eval_anatomy_runtime": 0.2679,
+      "eval_anatomy_samples_per_second": 7.464,
+      "eval_anatomy_steps_per_second": 3.732,
+      "step": 7200
+    },
+    {
+      "epoch": 2.9962546816479403,
+      "eval_college_mathematics_loss": 2.0544211864471436,
+      "eval_college_mathematics_runtime": 0.2676,
+      "eval_college_mathematics_samples_per_second": 7.475,
+      "eval_college_mathematics_steps_per_second": 3.737,
+      "step": 7200
+    },
+    {
+      "epoch": 2.9962546816479403,
+      "eval_international_law_loss": 3.067124843597412,
+      "eval_international_law_runtime": 0.267,
+      "eval_international_law_samples_per_second": 7.49,
+      "eval_international_law_steps_per_second": 3.745,
+      "step": 7200
+    },
+    {
+      "epoch": 3.0045776113191844,
+      "grad_norm": 0.3125,
+      "learning_rate": 4.807975298288149e-06,
+      "loss": 0.233,
+      "step": 7220
+    },
+    {
+      "epoch": 3.0129005409904286,
+      "grad_norm": 0.271484375,
+      "learning_rate": 4.804822498302021e-06,
+      "loss": 0.2302,
+      "step": 7240
+    },
+    {
+      "epoch": 3.0212234706616727,
+      "grad_norm": 0.244140625,
+      "learning_rate": 4.8016450771100455e-06,
+      "loss": 0.2335,
+      "step": 7260
+    },
+    {
+      "epoch": 3.0295464003329173,
+      "grad_norm": 0.26171875,
+      "learning_rate": 4.798443068654939e-06,
+      "loss": 0.2322,
+      "step": 7280
+    },
+    {
+      "epoch": 3.0378693300041615,
+      "grad_norm": 0.267578125,
+      "learning_rate": 4.795216507142074e-06,
+      "loss": 0.2298,
+      "step": 7300
+    },
+    {
+      "epoch": 3.0378693300041615,
+      "eval_main_loss": 0.23858603835105896,
+      "eval_main_runtime": 6.3548,
+      "eval_main_samples_per_second": 29.899,
+      "eval_main_steps_per_second": 3.777,
+      "step": 7300
+    },
+    {
+      "epoch": 3.0378693300041615,
+      "eval_anatomy_loss": 2.832512617111206,
+      "eval_anatomy_runtime": 0.2671,
+      "eval_anatomy_samples_per_second": 7.487,
+      "eval_anatomy_steps_per_second": 3.743,
+      "step": 7300
+    },
+    {
+      "epoch": 3.0378693300041615,
+      "eval_college_mathematics_loss": 2.0511765480041504,
+      "eval_college_mathematics_runtime": 0.2677,
+      "eval_college_mathematics_samples_per_second": 7.472,
+      "eval_college_mathematics_steps_per_second": 3.736,
+      "step": 7300
+    },
+    {
+      "epoch": 3.0378693300041615,
+      "eval_international_law_loss": 3.067012071609497,
+      "eval_international_law_runtime": 0.267,
+      "eval_international_law_samples_per_second": 7.491,
+      "eval_international_law_steps_per_second": 3.745,
+      "step": 7300
+    },
+    {
+      "epoch": 3.0461922596754056,
+      "grad_norm": 0.263671875,
+      "learning_rate": 4.791965427039109e-06,
+      "loss": 0.234,
+      "step": 7320
+    },
+    {
+      "epoch": 3.0545151893466502,
+      "grad_norm": 0.2734375,
+      "learning_rate": 4.788689863075622e-06,
+      "loss": 0.2319,
+      "step": 7340
+    },
+    {
+      "epoch": 3.0628381190178944,
+      "grad_norm": 0.31640625,
+      "learning_rate": 4.785389850242739e-06,
+      "loss": 0.2309,
+      "step": 7360
+    },
+    {
+      "epoch": 3.0711610486891385,
+      "grad_norm": 0.240234375,
+      "learning_rate": 4.78206542379276e-06,
+      "loss": 0.2283,
+      "step": 7380
+    },
+    {
+      "epoch": 3.0794839783603827,
+      "grad_norm": 0.279296875,
+      "learning_rate": 4.778716619238784e-06,
+      "loss": 0.23,
+      "step": 7400
+    },
+    {
+      "epoch": 3.0794839783603827,
+      "eval_main_loss": 0.23865698277950287,
+      "eval_main_runtime": 6.3366,
+      "eval_main_samples_per_second": 29.984,
+      "eval_main_steps_per_second": 3.788,
+      "step": 7400
+    },
+    {
+      "epoch": 3.0794839783603827,
+      "eval_anatomy_loss": 2.8329567909240723,
+      "eval_anatomy_runtime": 0.2666,
+      "eval_anatomy_samples_per_second": 7.502,
+      "eval_anatomy_steps_per_second": 3.751,
+      "step": 7400
+    },
+    {
+      "epoch": 3.0794839783603827,
+      "eval_college_mathematics_loss": 2.0505635738372803,
+      "eval_college_mathematics_runtime": 0.2663,
+      "eval_college_mathematics_samples_per_second": 7.509,
+      "eval_college_mathematics_steps_per_second": 3.755,
+      "step": 7400
+    },
+    {
+      "epoch": 3.0794839783603827,
+      "eval_international_law_loss": 3.06803035736084,
+      "eval_international_law_runtime": 0.2662,
+      "eval_international_law_samples_per_second": 7.512,
+      "eval_international_law_steps_per_second": 3.756,
+      "step": 7400
+    },
+    {
+      "epoch": 3.0878069080316273,
+      "grad_norm": 0.2412109375,
+      "learning_rate": 4.7753434723543266e-06,
+      "loss": 0.2308,
+      "step": 7420
+    },
+    {
+      "epoch": 3.0961298377028714,
+      "grad_norm": 0.28515625,
+      "learning_rate": 4.771946019172942e-06,
+      "loss": 0.2331,
+      "step": 7440
+    },
+    {
+      "epoch": 3.1044527673741156,
+      "grad_norm": 0.2890625,
+      "learning_rate": 4.768524295987835e-06,
+      "loss": 0.2383,
+      "step": 7460
+    },
+    {
+      "epoch": 3.11277569704536,
+      "grad_norm": 0.25,
+      "learning_rate": 4.765078339351472e-06,
+      "loss": 0.2327,
+      "step": 7480
+    },
+    {
+      "epoch": 3.1210986267166043,
+      "grad_norm": 0.3046875,
+      "learning_rate": 4.761608186075196e-06,
+      "loss": 0.2312,
+      "step": 7500
+    },
+    {
+      "epoch": 3.1210986267166043,
+      "eval_main_loss": 0.2385028600692749,
+      "eval_main_runtime": 6.3337,
+      "eval_main_samples_per_second": 29.998,
+      "eval_main_steps_per_second": 3.789,
+      "step": 7500
+    },
+    {
+      "epoch": 3.1210986267166043,
+      "eval_anatomy_loss": 2.834723949432373,
+      "eval_anatomy_runtime": 0.2659,
+      "eval_anatomy_samples_per_second": 7.522,
+      "eval_anatomy_steps_per_second": 3.761,
+      "step": 7500
+    },
+    {
+      "epoch": 3.1210986267166043,
+      "eval_college_mathematics_loss": 2.048158884048462,
+      "eval_college_mathematics_runtime": 0.2669,
+      "eval_college_mathematics_samples_per_second": 7.494,
+      "eval_college_mathematics_steps_per_second": 3.747,
+      "step": 7500
+    },
+    {
+      "epoch": 3.1210986267166043,
+      "eval_international_law_loss": 3.065969228744507,
+      "eval_international_law_runtime": 0.2652,
+      "eval_international_law_samples_per_second": 7.542,
+      "eval_international_law_steps_per_second": 3.771,
+      "step": 7500
+    },
+    {
+      "epoch": 3.1294215563878485,
+      "grad_norm": 0.30078125,
+      "learning_rate": 4.758113873228828e-06,
+      "loss": 0.2361,
+      "step": 7520
+    },
+    {
+      "epoch": 3.1377444860590926,
+      "grad_norm": 0.328125,
+      "learning_rate": 4.754595438140272e-06,
+      "loss": 0.2312,
+      "step": 7540
+    },
+    {
+      "epoch": 3.146067415730337,
+      "grad_norm": 0.265625,
+      "learning_rate": 4.75105291839512e-06,
+      "loss": 0.232,
+      "step": 7560
+    },
+    {
+      "epoch": 3.1543903454015814,
+      "grad_norm": 0.2353515625,
+      "learning_rate": 4.747486351836246e-06,
+      "loss": 0.2308,
+      "step": 7580
+    },
+    {
+      "epoch": 3.1627132750728255,
+      "grad_norm": 0.294921875,
+      "learning_rate": 4.743895776563403e-06,
+      "loss": 0.2334,
+      "step": 7600
+    },
+    {
+      "epoch": 3.1627132750728255,
+      "eval_main_loss": 0.23862504959106445,
+      "eval_main_runtime": 6.3275,
+      "eval_main_samples_per_second": 30.028,
+      "eval_main_steps_per_second": 3.793,
+      "step": 7600
+    },
+    {
+      "epoch": 3.1627132750728255,
+      "eval_anatomy_loss": 2.8308141231536865,
+      "eval_anatomy_runtime": 0.2657,
+      "eval_anatomy_samples_per_second": 7.528,
+      "eval_anatomy_steps_per_second": 3.764,
+      "step": 7600
+    },
+    {
+      "epoch": 3.1627132750728255,
+      "eval_college_mathematics_loss": 2.053100347518921,
+      "eval_college_mathematics_runtime": 0.2659,
+      "eval_college_mathematics_samples_per_second": 7.521,
+      "eval_college_mathematics_steps_per_second": 3.761,
+      "step": 7600
+    },
+    {
+      "epoch": 3.1627132750728255,
+      "eval_international_law_loss": 3.066927433013916,
+      "eval_international_law_runtime": 0.2659,
+      "eval_international_law_samples_per_second": 7.521,
+      "eval_international_law_steps_per_second": 3.76,
+      "step": 7600
+    },
+    {
+      "epoch": 3.17103620474407,
+      "grad_norm": 0.25390625,
+      "learning_rate": 4.740281230932817e-06,
+      "loss": 0.2345,
+      "step": 7620
+    },
+    {
+      "epoch": 3.1793591344153143,
+      "grad_norm": 0.236328125,
+      "learning_rate": 4.736642753556777e-06,
+      "loss": 0.2343,
+      "step": 7640
+    },
+    {
+      "epoch": 3.1876820640865584,
+      "grad_norm": 0.318359375,
+      "learning_rate": 4.732980383303223e-06,
+      "loss": 0.2327,
+      "step": 7660
+    },
+    {
+      "epoch": 3.1960049937578026,
+      "grad_norm": 0.30859375,
+      "learning_rate": 4.729294159295329e-06,
+      "loss": 0.2316,
+      "step": 7680
+    },
+    {
+      "epoch": 3.204327923429047,
+      "grad_norm": 0.310546875,
+      "learning_rate": 4.725584120911085e-06,
+      "loss": 0.2324,
+      "step": 7700
+    },
+    {
+      "epoch": 3.204327923429047,
+      "eval_main_loss": 0.2384970486164093,
+      "eval_main_runtime": 6.3308,
+      "eval_main_samples_per_second": 30.012,
+      "eval_main_steps_per_second": 3.791,
+      "step": 7700
+    },
+    {
+      "epoch": 3.204327923429047,
+      "eval_anatomy_loss": 2.8297290802001953,
+      "eval_anatomy_runtime": 0.2666,
+      "eval_anatomy_samples_per_second": 7.501,
+      "eval_anatomy_steps_per_second": 3.751,
+      "step": 7700
+    },
+    {
+      "epoch": 3.204327923429047,
+      "eval_college_mathematics_loss": 2.0536551475524902,
+      "eval_college_mathematics_runtime": 0.2648,
+      "eval_college_mathematics_samples_per_second": 7.552,
+      "eval_college_mathematics_steps_per_second": 3.776,
+      "step": 7700
+    },
+    {
+      "epoch": 3.204327923429047,
+      "eval_international_law_loss": 3.0671679973602295,
+      "eval_international_law_runtime": 0.2662,
+      "eval_international_law_samples_per_second": 7.513,
+      "eval_international_law_steps_per_second": 3.757,
+      "step": 7700
+    },
+    {
+      "epoch": 3.2126508531002913,
+      "grad_norm": 0.296875,
+      "learning_rate": 4.721850307782879e-06,
+      "loss": 0.2278,
+      "step": 7720
+    },
+    {
+      "epoch": 3.2209737827715355,
+      "grad_norm": 0.306640625,
+      "learning_rate": 4.718092759797073e-06,
+      "loss": 0.2338,
+      "step": 7740
+    },
+    {
+      "epoch": 3.22929671244278,
+      "grad_norm": 0.2734375,
+      "learning_rate": 4.714311517093573e-06,
+      "loss": 0.2315,
+      "step": 7760
+    },
+    {
+      "epoch": 3.237619642114024,
+      "grad_norm": 0.2412109375,
+      "learning_rate": 4.710506620065406e-06,
+      "loss": 0.2368,
+      "step": 7780
+    },
+    {
+      "epoch": 3.2459425717852683,
+      "grad_norm": 0.291015625,
+      "learning_rate": 4.706678109358285e-06,
+      "loss": 0.2386,
+      "step": 7800
+    },
+    {
+      "epoch": 3.2459425717852683,
+      "eval_main_loss": 0.23850186169147491,
+      "eval_main_runtime": 6.3235,
+      "eval_main_samples_per_second": 30.047,
+      "eval_main_steps_per_second": 3.795,
+      "step": 7800
+    },
+    {
+      "epoch": 3.2459425717852683,
+      "eval_anatomy_loss": 2.8328044414520264,
+      "eval_anatomy_runtime": 0.265,
+      "eval_anatomy_samples_per_second": 7.548,
+      "eval_anatomy_steps_per_second": 3.774,
+      "step": 7800
+    },
+    {
+      "epoch": 3.2459425717852683,
+      "eval_college_mathematics_loss": 2.048779249191284,
+      "eval_college_mathematics_runtime": 0.2654,
+      "eval_college_mathematics_samples_per_second": 7.535,
+      "eval_college_mathematics_steps_per_second": 3.768,
+      "step": 7800
+    },
+    {
+      "epoch": 3.2459425717852683,
+      "eval_international_law_loss": 3.0646255016326904,
+      "eval_international_law_runtime": 0.2657,
+      "eval_international_law_samples_per_second": 7.529,
+      "eval_international_law_steps_per_second": 3.764,
+      "step": 7800
+    },
+    {
+      "epoch": 3.2542655014565125,
+      "grad_norm": 0.2578125,
+      "learning_rate": 4.702826025870173e-06,
+      "loss": 0.2354,
+      "step": 7820
+    },
+    {
+      "epoch": 3.262588431127757,
+      "grad_norm": 0.29296875,
+      "learning_rate": 4.698950410750854e-06,
+      "loss": 0.231,
+      "step": 7840
+    },
+    {
+      "epoch": 3.2709113607990012,
+      "grad_norm": 0.275390625,
+      "learning_rate": 4.695051305401483e-06,
+      "loss": 0.232,
+      "step": 7860
+    },
+    {
+      "epoch": 3.2792342904702454,
+      "grad_norm": 0.341796875,
+      "learning_rate": 4.691128751474149e-06,
+      "loss": 0.2307,
+      "step": 7880
+    },
+    {
+      "epoch": 3.28755722014149,
+      "grad_norm": 0.3046875,
+      "learning_rate": 4.6871827908714345e-06,
+      "loss": 0.2306,
+      "step": 7900
+    },
+    {
+      "epoch": 3.28755722014149,
+      "eval_main_loss": 0.23848366737365723,
+      "eval_main_runtime": 6.3519,
+      "eval_main_samples_per_second": 29.912,
+      "eval_main_steps_per_second": 3.778,
+      "step": 7900
+    },
+    {
+      "epoch": 3.28755722014149,
+      "eval_anatomy_loss": 2.834224224090576,
+      "eval_anatomy_runtime": 0.2672,
+      "eval_anatomy_samples_per_second": 7.485,
+      "eval_anatomy_steps_per_second": 3.742,
+      "step": 7900
+    },
+    {
+      "epoch": 3.28755722014149,
+      "eval_college_mathematics_loss": 2.05056095123291,
+      "eval_college_mathematics_runtime": 0.2672,
+      "eval_college_mathematics_samples_per_second": 7.486,
+      "eval_college_mathematics_steps_per_second": 3.743,
+      "step": 7900
+    },
+    {
+      "epoch": 3.28755722014149,
+      "eval_international_law_loss": 3.0639660358428955,
+      "eval_international_law_runtime": 0.2668,
+      "eval_international_law_samples_per_second": 7.496,
+      "eval_international_law_steps_per_second": 3.748,
+      "step": 7900
+    },
+    {
+      "epoch": 3.295880149812734,
+      "grad_norm": 0.291015625,
+      "learning_rate": 4.6832134657459586e-06,
+      "loss": 0.2349,
+      "step": 7920
+    },
+    {
+      "epoch": 3.3042030794839783,
+      "grad_norm": 0.291015625,
+      "learning_rate": 4.679220818499932e-06,
+      "loss": 0.2316,
+      "step": 7940
+    },
+    {
+      "epoch": 3.3125260091552224,
+      "grad_norm": 0.236328125,
+      "learning_rate": 4.675204891784706e-06,
+      "loss": 0.2326,
+      "step": 7960
+    },
+    {
+      "epoch": 3.320848938826467,
+      "grad_norm": 0.259765625,
+      "learning_rate": 4.671165728500311e-06,
+      "loss": 0.2325,
+      "step": 7980
+    },
+    {
+      "epoch": 3.329171868497711,
+      "grad_norm": 0.306640625,
+      "learning_rate": 4.667103371795003e-06,
+      "loss": 0.2305,
+      "step": 8000
+    },
+    {
+      "epoch": 3.329171868497711,
+      "eval_main_loss": 0.23847457766532898,
+      "eval_main_runtime": 6.3564,
+      "eval_main_samples_per_second": 29.891,
+      "eval_main_steps_per_second": 3.776,
+      "step": 8000
+    },
+    {
+      "epoch": 3.329171868497711,
+      "eval_anatomy_loss": 2.8327555656433105,
+      "eval_anatomy_runtime": 0.2663,
+      "eval_anatomy_samples_per_second": 7.509,
+      "eval_anatomy_steps_per_second": 3.755,
+      "step": 8000
+    },
+    {
+      "epoch": 3.329171868497711,
+      "eval_college_mathematics_loss": 2.0528974533081055,
+      "eval_college_mathematics_runtime": 0.2665,
+      "eval_college_mathematics_samples_per_second": 7.505,
+      "eval_college_mathematics_steps_per_second": 3.753,
+      "step": 8000
+    },
+    {
+      "epoch": 3.329171868497711,
+      "eval_international_law_loss": 3.065342903137207,
+      "eval_international_law_runtime": 0.2694,
+      "eval_international_law_samples_per_second": 7.423,
+      "eval_international_law_steps_per_second": 3.712,
+      "step": 8000
+    },
+    {
+      "epoch": 3.3374947981689553,
+      "grad_norm": 0.3203125,
+      "learning_rate": 4.6630178650648005e-06,
+      "loss": 0.226,
+      "step": 8020
+    },
+    {
+      "epoch": 3.3458177278402,
+      "grad_norm": 0.28515625,
+      "learning_rate": 4.658909251953023e-06,
+      "loss": 0.2341,
+      "step": 8040
+    },
+    {
+      "epoch": 3.354140657511444,
+      "grad_norm": 0.2578125,
+      "learning_rate": 4.654777576349822e-06,
+      "loss": 0.232,
+      "step": 8060
+    },
+    {
+      "epoch": 3.3624635871826882,
+      "grad_norm": 0.271484375,
+      "learning_rate": 4.650622882391713e-06,
+      "loss": 0.2266,
+      "step": 8080
+    },
+    {
+      "epoch": 3.370786516853933,
+      "grad_norm": 0.271484375,
+      "learning_rate": 4.646445214461105e-06,
+      "loss": 0.2328,
+      "step": 8100
+    },
+    {
+      "epoch": 3.370786516853933,
+      "eval_main_loss": 0.23838673532009125,
+      "eval_main_runtime": 6.3504,
+      "eval_main_samples_per_second": 29.919,
+      "eval_main_steps_per_second": 3.779,
+      "step": 8100
+    },
+    {
+      "epoch": 3.370786516853933,
+      "eval_anatomy_loss": 2.8322196006774902,
+      "eval_anatomy_runtime": 0.266,
+      "eval_anatomy_samples_per_second": 7.518,
+      "eval_anatomy_steps_per_second": 3.759,
+      "step": 8100
+    },
+    {
+      "epoch": 3.370786516853933,
+      "eval_college_mathematics_loss": 2.0516698360443115,
+      "eval_college_mathematics_runtime": 0.2667,
+      "eval_college_mathematics_samples_per_second": 7.499,
+      "eval_college_mathematics_steps_per_second": 3.75,
+      "step": 8100
+    },
+    {
+      "epoch": 3.370786516853933,
+      "eval_international_law_loss": 3.0663843154907227,
+      "eval_international_law_runtime": 0.2677,
+      "eval_international_law_samples_per_second": 7.47,
+      "eval_international_law_steps_per_second": 3.735,
+      "step": 8100
+    },
+    {
+      "epoch": 3.379109446525177,
+      "grad_norm": 0.30078125,
+      "learning_rate": 4.642244617185827e-06,
+      "loss": 0.231,
+      "step": 8120
+    },
+    {
+      "epoch": 3.387432376196421,
+      "grad_norm": 0.267578125,
+      "learning_rate": 4.6380211354386475e-06,
+      "loss": 0.2312,
+      "step": 8140
+    },
+    {
+      "epoch": 3.3957553058676653,
+      "grad_norm": 0.263671875,
+      "learning_rate": 4.633774814336801e-06,
+      "loss": 0.2321,
+      "step": 8160
+    },
+    {
+      "epoch": 3.40407823553891,
+      "grad_norm": 0.333984375,
+      "learning_rate": 4.6295056992415026e-06,
+      "loss": 0.2298,
+      "step": 8180
+    },
+    {
+      "epoch": 3.412401165210154,
+      "grad_norm": 0.3125,
+      "learning_rate": 4.625213835757458e-06,
+      "loss": 0.2348,
+      "step": 8200
+    },
+    {
+      "epoch": 3.412401165210154,
+      "eval_main_loss": 0.23835770785808563,
+      "eval_main_runtime": 6.3288,
+      "eval_main_samples_per_second": 30.022,
+      "eval_main_steps_per_second": 3.792,
+      "step": 8200
+    },
+    {
+      "epoch": 3.412401165210154,
+      "eval_anatomy_loss": 2.8323183059692383,
+      "eval_anatomy_runtime": 0.2656,
+      "eval_anatomy_samples_per_second": 7.531,
+      "eval_anatomy_steps_per_second": 3.766,
+      "step": 8200
+    },
+    {
+      "epoch": 3.412401165210154,
+      "eval_college_mathematics_loss": 2.0520193576812744,
+      "eval_college_mathematics_runtime": 0.2676,
+      "eval_college_mathematics_samples_per_second": 7.474,
+      "eval_college_mathematics_steps_per_second": 3.737,
+      "step": 8200
+    },
+    {
+      "epoch": 3.412401165210154,
+      "eval_international_law_loss": 3.0666942596435547,
+      "eval_international_law_runtime": 0.2666,
+      "eval_international_law_samples_per_second": 7.502,
+      "eval_international_law_steps_per_second": 3.751,
+      "step": 8200
+    },
+    {
+      "epoch": 3.420724094881398,
+      "grad_norm": 0.263671875,
+      "learning_rate": 4.62089926973239e-06,
+      "loss": 0.2312,
+      "step": 8220
+    },
+    {
+      "epoch": 3.4290470245526423,
+      "grad_norm": 0.267578125,
+      "learning_rate": 4.616562047256536e-06,
+      "loss": 0.23,
+      "step": 8240
+    },
+    {
+      "epoch": 3.437369954223887,
+      "grad_norm": 0.28125,
+      "learning_rate": 4.612202214662161e-06,
+      "loss": 0.2325,
+      "step": 8260
+    },
+    {
+      "epoch": 3.445692883895131,
+      "grad_norm": 0.3046875,
+      "learning_rate": 4.6078198185230605e-06,
+      "loss": 0.2323,
+      "step": 8280
+    },
+    {
+      "epoch": 3.454015813566375,
+      "grad_norm": 0.28125,
+      "learning_rate": 4.603414905654069e-06,
+      "loss": 0.2306,
+      "step": 8300
+    },
+    {
+      "epoch": 3.454015813566375,
+      "eval_main_loss": 0.23845936357975006,
+      "eval_main_runtime": 6.3281,
+      "eval_main_samples_per_second": 30.025,
+      "eval_main_steps_per_second": 3.793,
+      "step": 8300
+    },
+    {
+      "epoch": 3.454015813566375,
+      "eval_anatomy_loss": 2.8332934379577637,
+      "eval_anatomy_runtime": 0.2651,
+      "eval_anatomy_samples_per_second": 7.543,
+      "eval_anatomy_steps_per_second": 3.772,
+      "step": 8300
+    },
+    {
+      "epoch": 3.454015813566375,
+      "eval_college_mathematics_loss": 2.0525290966033936,
+      "eval_college_mathematics_runtime": 0.2658,
+      "eval_college_mathematics_samples_per_second": 7.524,
+      "eval_college_mathematics_steps_per_second": 3.762,
+      "step": 8300
+    },
+    {
+      "epoch": 3.454015813566375,
+      "eval_international_law_loss": 3.0639939308166504,
+      "eval_international_law_runtime": 0.2664,
+      "eval_international_law_samples_per_second": 7.508,
+      "eval_international_law_steps_per_second": 3.754,
+      "step": 8300
+    },
+    {
+      "epoch": 3.46233874323762,
+      "grad_norm": 0.2734375,
+      "learning_rate": 4.5989875231105514e-06,
+      "loss": 0.2328,
+      "step": 8320
+    },
+    {
+      "epoch": 3.470661672908864,
+      "grad_norm": 0.25390625,
+      "learning_rate": 4.594537718187906e-06,
+      "loss": 0.2274,
+      "step": 8340
+    },
+    {
+      "epoch": 3.478984602580108,
+      "grad_norm": 0.287109375,
+      "learning_rate": 4.590065538421056e-06,
+      "loss": 0.2304,
+      "step": 8360
+    },
+    {
+      "epoch": 3.4873075322513527,
+      "grad_norm": 0.255859375,
+      "learning_rate": 4.585571031583946e-06,
+      "loss": 0.2292,
+      "step": 8380
+    },
+    {
+      "epoch": 3.495630461922597,
+      "grad_norm": 0.2734375,
+      "learning_rate": 4.581054245689026e-06,
+      "loss": 0.2309,
+      "step": 8400
+    },
+    {
+      "epoch": 3.495630461922597,
+      "eval_main_loss": 0.23835162818431854,
+      "eval_main_runtime": 6.3575,
+      "eval_main_samples_per_second": 29.886,
+      "eval_main_steps_per_second": 3.775,
+      "step": 8400
+    },
+    {
+      "epoch": 3.495630461922597,
+      "eval_anatomy_loss": 2.8338778018951416,
+      "eval_anatomy_runtime": 0.2661,
+      "eval_anatomy_samples_per_second": 7.516,
+      "eval_anatomy_steps_per_second": 3.758,
+      "step": 8400
+    },
+    {
+      "epoch": 3.495630461922597,
+      "eval_college_mathematics_loss": 2.053218126296997,
+      "eval_college_mathematics_runtime": 0.2688,
+      "eval_college_mathematics_samples_per_second": 7.442,
+      "eval_college_mathematics_steps_per_second": 3.721,
+      "step": 8400
+    },
+    {
+      "epoch": 3.495630461922597,
+      "eval_international_law_loss": 3.065638303756714,
+      "eval_international_law_runtime": 0.2664,
+      "eval_international_law_samples_per_second": 7.508,
+      "eval_international_law_steps_per_second": 3.754,
+      "step": 8400
+    },
+    {
+      "epoch": 3.503953391593841,
+      "grad_norm": 0.30859375,
+      "learning_rate": 4.576515228986743e-06,
+      "loss": 0.235,
+      "step": 8420
+    },
+    {
+      "epoch": 3.512276321265085,
+      "grad_norm": 0.328125,
+      "learning_rate": 4.571954029965024e-06,
+      "loss": 0.2322,
+      "step": 8440
+    },
+    {
+      "epoch": 3.5205992509363297,
+      "grad_norm": 0.322265625,
+      "learning_rate": 4.567370697348759e-06,
+      "loss": 0.2353,
+      "step": 8460
+    },
+    {
+      "epoch": 3.528922180607574,
+      "grad_norm": 0.27734375,
+      "learning_rate": 4.5627652800992765e-06,
+      "loss": 0.2312,
+      "step": 8480
+    },
+    {
+      "epoch": 3.537245110278818,
+      "grad_norm": 0.283203125,
+      "learning_rate": 4.558137827413825e-06,
+      "loss": 0.2317,
+      "step": 8500
+    },
+    {
+      "epoch": 3.537245110278818,
+      "eval_main_loss": 0.23841607570648193,
+      "eval_main_runtime": 6.3538,
+      "eval_main_samples_per_second": 29.903,
+      "eval_main_steps_per_second": 3.777,
+      "step": 8500
+    },
+    {
+      "epoch": 3.537245110278818,
+      "eval_anatomy_loss": 2.83097505569458,
+      "eval_anatomy_runtime": 0.2657,
+      "eval_anatomy_samples_per_second": 7.528,
+      "eval_anatomy_steps_per_second": 3.764,
+      "step": 8500
+    },
+    {
+      "epoch": 3.537245110278818,
+      "eval_college_mathematics_loss": 2.0496573448181152,
+      "eval_college_mathematics_runtime": 0.2674,
+      "eval_college_mathematics_samples_per_second": 7.48,
+      "eval_college_mathematics_steps_per_second": 3.74,
+      "step": 8500
+    },
+    {
+      "epoch": 3.537245110278818,
+      "eval_international_law_loss": 3.065972089767456,
+      "eval_international_law_runtime": 0.2675,
+      "eval_international_law_samples_per_second": 7.475,
+      "eval_international_law_steps_per_second": 3.738,
+      "step": 8500
+    },
+    {
+      "epoch": 3.545568039950062,
+      "grad_norm": 0.28515625,
+      "learning_rate": 4.5534883887250495e-06,
+      "loss": 0.235,
+      "step": 8520
+    },
+    {
+      "epoch": 3.553890969621307,
+      "grad_norm": 0.28125,
+      "learning_rate": 4.548817013700454e-06,
+      "loss": 0.2341,
+      "step": 8540
+    },
+    {
+      "epoch": 3.562213899292551,
+      "grad_norm": 0.294921875,
+      "learning_rate": 4.5441237522418804e-06,
+      "loss": 0.2295,
+      "step": 8560
+    },
+    {
+      "epoch": 3.570536828963795,
+      "grad_norm": 0.302734375,
+      "learning_rate": 4.53940865448497e-06,
+      "loss": 0.2318,
+      "step": 8580
+    },
+    {
+      "epoch": 3.5788597586350397,
+      "grad_norm": 0.30078125,
+      "learning_rate": 4.534671770798633e-06,
+      "loss": 0.2298,
+      "step": 8600
+    },
+    {
+      "epoch": 3.5788597586350397,
+      "eval_main_loss": 0.23844130337238312,
+      "eval_main_runtime": 6.3401,
+      "eval_main_samples_per_second": 29.968,
+      "eval_main_steps_per_second": 3.785,
+      "step": 8600
+    },
+    {
+      "epoch": 3.5788597586350397,
+      "eval_anatomy_loss": 2.8311820030212402,
+      "eval_anatomy_runtime": 0.2673,
+      "eval_anatomy_samples_per_second": 7.482,
+      "eval_anatomy_steps_per_second": 3.741,
+      "step": 8600
+    },
+    {
+      "epoch": 3.5788597586350397,
+      "eval_college_mathematics_loss": 2.051837921142578,
+      "eval_college_mathematics_runtime": 0.2661,
+      "eval_college_mathematics_samples_per_second": 7.517,
+      "eval_college_mathematics_steps_per_second": 3.758,
+      "step": 8600
+    },
+    {
+      "epoch": 3.5788597586350397,
+      "eval_international_law_loss": 3.0655837059020996,
+      "eval_international_law_runtime": 0.2666,
+      "eval_international_law_samples_per_second": 7.503,
+      "eval_international_law_steps_per_second": 3.751,
+      "step": 8600
+    },
+    {
+      "epoch": 3.587182688306284,
+      "grad_norm": 0.26953125,
+      "learning_rate": 4.529913151784504e-06,
+      "loss": 0.2311,
+      "step": 8620
+    },
+    {
+      "epoch": 3.595505617977528,
+      "grad_norm": 0.294921875,
+      "learning_rate": 4.525132848276405e-06,
+      "loss": 0.2341,
+      "step": 8640
+    },
+    {
+      "epoch": 3.6038285476487726,
+      "grad_norm": 0.296875,
+      "learning_rate": 4.520330911339805e-06,
+      "loss": 0.2343,
+      "step": 8660
+    },
+    {
+      "epoch": 3.6121514773200167,
+      "grad_norm": 0.28515625,
+      "learning_rate": 4.5155073922712665e-06,
+      "loss": 0.2324,
+      "step": 8680
+    },
+    {
+      "epoch": 3.620474406991261,
+      "grad_norm": 0.2119140625,
+      "learning_rate": 4.510662342597907e-06,
+      "loss": 0.2335,
+      "step": 8700
+    },
+    {
+      "epoch": 3.620474406991261,
+      "eval_main_loss": 0.23842211067676544,
+      "eval_main_runtime": 6.3337,
+      "eval_main_samples_per_second": 29.998,
+      "eval_main_steps_per_second": 3.789,
+      "step": 8700
+    },
+    {
+      "epoch": 3.620474406991261,
+      "eval_anatomy_loss": 2.8317272663116455,
+      "eval_anatomy_runtime": 0.2656,
+      "eval_anatomy_samples_per_second": 7.531,
+      "eval_anatomy_steps_per_second": 3.765,
+      "step": 8700
+    },
+    {
+      "epoch": 3.620474406991261,
+      "eval_college_mathematics_loss": 2.0518221855163574,
+      "eval_college_mathematics_runtime": 0.2661,
+      "eval_college_mathematics_samples_per_second": 7.515,
+      "eval_college_mathematics_steps_per_second": 3.758,
+      "step": 8700
+    },
+    {
+      "epoch": 3.620474406991261,
+      "eval_international_law_loss": 3.06583571434021,
+      "eval_international_law_runtime": 0.2669,
+      "eval_international_law_samples_per_second": 7.492,
+      "eval_international_law_steps_per_second": 3.746,
+      "step": 8700
+    },
+    {
+      "epoch": 3.628797336662505,
+      "grad_norm": 0.263671875,
+      "learning_rate": 4.505795814076842e-06,
+      "loss": 0.2318,
+      "step": 8720
+    },
+    {
+      "epoch": 3.6371202663337496,
+      "grad_norm": 0.2578125,
+      "learning_rate": 4.5009078586946355e-06,
+      "loss": 0.2333,
+      "step": 8740
+    },
+    {
+      "epoch": 3.645443196004994,
+      "grad_norm": 0.279296875,
+      "learning_rate": 4.495998528666741e-06,
+      "loss": 0.2311,
+      "step": 8760
+    },
+    {
+      "epoch": 3.653766125676238,
+      "grad_norm": 0.259765625,
+      "learning_rate": 4.491067876436949e-06,
+      "loss": 0.2322,
+      "step": 8780
+    },
+    {
+      "epoch": 3.662089055347482,
+      "grad_norm": 0.318359375,
+      "learning_rate": 4.486115954676821e-06,
+      "loss": 0.2342,
+      "step": 8800
+    },
+    {
+      "epoch": 3.662089055347482,
+      "eval_main_loss": 0.23842394351959229,
+      "eval_main_runtime": 6.3208,
+      "eval_main_samples_per_second": 30.06,
+      "eval_main_steps_per_second": 3.797,
+      "step": 8800
+    },
+    {
+      "epoch": 3.662089055347482,
+      "eval_anatomy_loss": 2.8346548080444336,
+      "eval_anatomy_runtime": 0.2658,
+      "eval_anatomy_samples_per_second": 7.526,
+      "eval_anatomy_steps_per_second": 3.763,
+      "step": 8800
+    },
+    {
+      "epoch": 3.662089055347482,
+      "eval_college_mathematics_loss": 2.0510027408599854,
+      "eval_college_mathematics_runtime": 0.2657,
+      "eval_college_mathematics_samples_per_second": 7.526,
+      "eval_college_mathematics_steps_per_second": 3.763,
+      "step": 8800
+    },
+    {
+      "epoch": 3.662089055347482,
+      "eval_international_law_loss": 3.0672590732574463,
+      "eval_international_law_runtime": 0.2669,
+      "eval_international_law_samples_per_second": 7.493,
+      "eval_international_law_steps_per_second": 3.747,
+      "step": 8800
+    },
+    {
+      "epoch": 3.6704119850187267,
+      "grad_norm": 0.25390625,
+      "learning_rate": 4.481142816285133e-06,
+      "loss": 0.2331,
+      "step": 8820
+    },
+    {
+      "epoch": 3.678734914689971,
+      "grad_norm": 0.2275390625,
+      "learning_rate": 4.476148514387305e-06,
+      "loss": 0.2343,
+      "step": 8840
+    },
+    {
+      "epoch": 3.687057844361215,
+      "grad_norm": 0.267578125,
+      "learning_rate": 4.471133102334836e-06,
+      "loss": 0.2323,
+      "step": 8860
+    },
+    {
+      "epoch": 3.6953807740324596,
+      "grad_norm": 0.310546875,
+      "learning_rate": 4.4660966337047325e-06,
+      "loss": 0.2326,
+      "step": 8880
+    },
+    {
+      "epoch": 3.7037037037037037,
+      "grad_norm": 0.26953125,
+      "learning_rate": 4.46103916229894e-06,
+      "loss": 0.235,
+      "step": 8900
+    },
+    {
+      "epoch": 3.7037037037037037,
+      "eval_main_loss": 0.2383807897567749,
+      "eval_main_runtime": 6.326,
+      "eval_main_samples_per_second": 30.035,
+      "eval_main_steps_per_second": 3.794,
+      "step": 8900
+    },
+    {
+      "epoch": 3.7037037037037037,
+      "eval_anatomy_loss": 2.830998420715332,
+      "eval_anatomy_runtime": 0.2656,
+      "eval_anatomy_samples_per_second": 7.531,
+      "eval_anatomy_steps_per_second": 3.765,
+      "step": 8900
+    },
+    {
+      "epoch": 3.7037037037037037,
+      "eval_college_mathematics_loss": 2.051891803741455,
+      "eval_college_mathematics_runtime": 0.2656,
+      "eval_college_mathematics_samples_per_second": 7.53,
+      "eval_college_mathematics_steps_per_second": 3.765,
+      "step": 8900
+    },
+    {
+      "epoch": 3.7037037037037037,
+      "eval_international_law_loss": 3.0654711723327637,
+      "eval_international_law_runtime": 0.2652,
+      "eval_international_law_samples_per_second": 7.541,
+      "eval_international_law_steps_per_second": 3.77,
+      "step": 8900
+    },
+    {
+      "epoch": 3.712026633374948,
+      "grad_norm": 0.26171875,
+      "learning_rate": 4.455960742143762e-06,
+      "loss": 0.2316,
+      "step": 8920
+    },
+    {
+      "epoch": 3.7203495630461925,
+      "grad_norm": 0.2373046875,
+      "learning_rate": 4.450861427489291e-06,
+      "loss": 0.2319,
+      "step": 8940
+    },
+    {
+      "epoch": 3.7286724927174366,
+      "grad_norm": 0.248046875,
+      "learning_rate": 4.445741272808821e-06,
+      "loss": 0.2296,
+      "step": 8960
+    },
+    {
+      "epoch": 3.7369954223886808,
+      "grad_norm": 0.2275390625,
+      "learning_rate": 4.440600332798269e-06,
+      "loss": 0.2317,
+      "step": 8980
+    },
+    {
+      "epoch": 3.7453183520599254,
+      "grad_norm": 0.28515625,
+      "learning_rate": 4.435438662375593e-06,
+      "loss": 0.237,
+      "step": 9000
+    },
+    {
+      "epoch": 3.7453183520599254,
+      "eval_main_loss": 0.23835314810276031,
+      "eval_main_runtime": 6.3284,
+      "eval_main_samples_per_second": 30.023,
+      "eval_main_steps_per_second": 3.792,
+      "step": 9000
+    },
+    {
+      "epoch": 3.7453183520599254,
+      "eval_anatomy_loss": 2.832582473754883,
+      "eval_anatomy_runtime": 0.2663,
+      "eval_anatomy_samples_per_second": 7.509,
+      "eval_anatomy_steps_per_second": 3.755,
+      "step": 9000
+    },
+    {
+      "epoch": 3.7453183520599254,
+      "eval_college_mathematics_loss": 2.0516104698181152,
+      "eval_college_mathematics_runtime": 0.2657,
+      "eval_college_mathematics_samples_per_second": 7.526,
+      "eval_college_mathematics_steps_per_second": 3.763,
+      "step": 9000
+    },
+    {
+      "epoch": 3.7453183520599254,
+      "eval_international_law_loss": 3.0663774013519287,
+      "eval_international_law_runtime": 0.2661,
+      "eval_international_law_samples_per_second": 7.517,
+      "eval_international_law_steps_per_second": 3.758,
+      "step": 9000
+    },
+    {
+      "epoch": 3.7536412817311695,
+      "grad_norm": 0.2412109375,
+      "learning_rate": 4.430256316680201e-06,
+      "loss": 0.2296,
+      "step": 9020
+    },
+    {
+      "epoch": 3.7619642114024137,
+      "grad_norm": 0.248046875,
+      "learning_rate": 4.425053351072365e-06,
+      "loss": 0.2348,
+      "step": 9040
+    },
+    {
+      "epoch": 3.770287141073658,
+      "grad_norm": 0.294921875,
+      "learning_rate": 4.419829821132629e-06,
+      "loss": 0.2318,
+      "step": 9060
+    },
+    {
+      "epoch": 3.778610070744902,
+      "grad_norm": 0.306640625,
+      "learning_rate": 4.414585782661215e-06,
+      "loss": 0.2311,
+      "step": 9080
+    },
+    {
+      "epoch": 3.7869330004161466,
+      "grad_norm": 0.2890625,
+      "learning_rate": 4.4093212916774245e-06,
+      "loss": 0.2342,
+      "step": 9100
+    },
+    {
+      "epoch": 3.7869330004161466,
+      "eval_main_loss": 0.23832525312900543,
+      "eval_main_runtime": 6.3321,
+      "eval_main_samples_per_second": 30.006,
+      "eval_main_steps_per_second": 3.79,
+      "step": 9100
+    },
+    {
+      "epoch": 3.7869330004161466,
+      "eval_anatomy_loss": 2.83284854888916,
+      "eval_anatomy_runtime": 0.2678,
+      "eval_anatomy_samples_per_second": 7.468,
+      "eval_anatomy_steps_per_second": 3.734,
+      "step": 9100
+    },
+    {
+      "epoch": 3.7869330004161466,
+      "eval_college_mathematics_loss": 2.0502493381500244,
+      "eval_college_mathematics_runtime": 0.2658,
+      "eval_college_mathematics_samples_per_second": 7.526,
+      "eval_college_mathematics_steps_per_second": 3.763,
+      "step": 9100
+    },
+    {
+      "epoch": 3.7869330004161466,
+      "eval_international_law_loss": 3.0675525665283203,
+      "eval_international_law_runtime": 0.2664,
+      "eval_international_law_samples_per_second": 7.507,
+      "eval_international_law_steps_per_second": 3.754,
+      "step": 9100
+    },
+    {
+      "epoch": 3.7952559300873907,
+      "grad_norm": 0.279296875,
+      "learning_rate": 4.404036404419045e-06,
+      "loss": 0.2319,
+      "step": 9120
+    },
+    {
+      "epoch": 3.803578859758635,
+      "grad_norm": 0.314453125,
+      "learning_rate": 4.398731177341747e-06,
+      "loss": 0.2277,
+      "step": 9140
+    },
+    {
+      "epoch": 3.8119017894298795,
+      "grad_norm": 0.2734375,
+      "learning_rate": 4.39340566711848e-06,
+      "loss": 0.2322,
+      "step": 9160
+    },
+    {
+      "epoch": 3.8202247191011236,
+      "grad_norm": 0.279296875,
+      "learning_rate": 4.388059930638865e-06,
+      "loss": 0.2358,
+      "step": 9180
+    },
+    {
+      "epoch": 3.8285476487723678,
+      "grad_norm": 0.271484375,
+      "learning_rate": 4.3826940250085925e-06,
+      "loss": 0.2254,
+      "step": 9200
+    },
+    {
+      "epoch": 3.8285476487723678,
+      "eval_main_loss": 0.2383318841457367,
+      "eval_main_runtime": 6.3299,
+      "eval_main_samples_per_second": 30.016,
+      "eval_main_steps_per_second": 3.792,
+      "step": 9200
+    },
+    {
+      "epoch": 3.8285476487723678,
+      "eval_anatomy_loss": 2.833568811416626,
+      "eval_anatomy_runtime": 0.2667,
+      "eval_anatomy_samples_per_second": 7.5,
+      "eval_anatomy_steps_per_second": 3.75,
+      "step": 9200
+    },
+    {
+      "epoch": 3.8285476487723678,
+      "eval_college_mathematics_loss": 2.055227756500244,
+      "eval_college_mathematics_runtime": 0.2648,
+      "eval_college_mathematics_samples_per_second": 7.554,
+      "eval_college_mathematics_steps_per_second": 3.777,
+      "step": 9200
+    },
+    {
+      "epoch": 3.8285476487723678,
+      "eval_international_law_loss": 3.0670459270477295,
+      "eval_international_law_runtime": 0.2661,
+      "eval_international_law_samples_per_second": 7.516,
+      "eval_international_law_steps_per_second": 3.758,
+      "step": 9200
+    },
+    {
+      "epoch": 3.8368705784436123,
+      "grad_norm": 0.298828125,
+      "learning_rate": 4.377308007548809e-06,
+      "loss": 0.2281,
+      "step": 9220
+    },
+    {
+      "epoch": 3.8451935081148565,
+      "grad_norm": 0.26171875,
+      "learning_rate": 4.371901935795504e-06,
+      "loss": 0.2323,
+      "step": 9240
+    },
+    {
+      "epoch": 3.8535164377861006,
+      "grad_norm": 0.2119140625,
+      "learning_rate": 4.3664758674988984e-06,
+      "loss": 0.228,
+      "step": 9260
+    },
+    {
+      "epoch": 3.8618393674573452,
+      "grad_norm": 0.2490234375,
+      "learning_rate": 4.361029860622822e-06,
+      "loss": 0.2315,
+      "step": 9280
+    },
+    {
+      "epoch": 3.8701622971285894,
+      "grad_norm": 0.263671875,
+      "learning_rate": 4.355563973344104e-06,
+      "loss": 0.231,
+      "step": 9300
+    },
+    {
+      "epoch": 3.8701622971285894,
+      "eval_main_loss": 0.23839576542377472,
+      "eval_main_runtime": 6.3285,
+      "eval_main_samples_per_second": 30.023,
+      "eval_main_steps_per_second": 3.792,
+      "step": 9300
+    },
+    {
+      "epoch": 3.8701622971285894,
+      "eval_anatomy_loss": 2.8333206176757812,
+      "eval_anatomy_runtime": 0.2661,
+      "eval_anatomy_samples_per_second": 7.516,
+      "eval_anatomy_steps_per_second": 3.758,
+      "step": 9300
+    },
+    {
+      "epoch": 3.8701622971285894,
+      "eval_college_mathematics_loss": 2.048659563064575,
+      "eval_college_mathematics_runtime": 0.2666,
+      "eval_college_mathematics_samples_per_second": 7.5,
+      "eval_college_mathematics_steps_per_second": 3.75,
+      "step": 9300
+    },
+    {
+      "epoch": 3.8701622971285894,
+      "eval_international_law_loss": 3.0655484199523926,
+      "eval_international_law_runtime": 0.2664,
+      "eval_international_law_samples_per_second": 7.508,
+      "eval_international_law_steps_per_second": 3.754,
+      "step": 9300
+    },
+    {
+      "epoch": 3.8784852267998335,
+      "grad_norm": 0.2373046875,
+      "learning_rate": 4.3500782640519375e-06,
+      "loss": 0.2261,
+      "step": 9320
+    },
+    {
+      "epoch": 3.8868081564710777,
+      "grad_norm": 0.294921875,
+      "learning_rate": 4.344572791347272e-06,
+      "loss": 0.2274,
+      "step": 9340
+    },
+    {
+      "epoch": 3.895131086142322,
+      "grad_norm": 0.302734375,
+      "learning_rate": 4.339047614042172e-06,
+      "loss": 0.2316,
+      "step": 9360
+    },
+    {
+      "epoch": 3.9034540158135664,
+      "grad_norm": 0.33203125,
+      "learning_rate": 4.3335027911592004e-06,
+      "loss": 0.23,
+      "step": 9380
+    },
+    {
+      "epoch": 3.9117769454848106,
+      "grad_norm": 0.30078125,
+      "learning_rate": 4.327938381930782e-06,
+      "loss": 0.2318,
+      "step": 9400
+    },
+    {
+      "epoch": 3.9117769454848106,
+      "eval_main_loss": 0.23833923041820526,
+      "eval_main_runtime": 6.3193,
+      "eval_main_samples_per_second": 30.067,
+      "eval_main_steps_per_second": 3.798,
+      "step": 9400
+    },
+    {
+      "epoch": 3.9117769454848106,
+      "eval_anatomy_loss": 2.8319694995880127,
+      "eval_anatomy_runtime": 0.266,
+      "eval_anatomy_samples_per_second": 7.52,
+      "eval_anatomy_steps_per_second": 3.76,
+      "step": 9400
+    },
+    {
+      "epoch": 3.9117769454848106,
+      "eval_college_mathematics_loss": 2.0500192642211914,
+      "eval_college_mathematics_runtime": 0.2666,
+      "eval_college_mathematics_samples_per_second": 7.502,
+      "eval_college_mathematics_steps_per_second": 3.751,
+      "step": 9400
+    },
+    {
+      "epoch": 3.9117769454848106,
+      "eval_international_law_loss": 3.0644402503967285,
+      "eval_international_law_runtime": 0.2672,
+      "eval_international_law_samples_per_second": 7.486,
+      "eval_international_law_steps_per_second": 3.743,
+      "step": 9400
+    },
+    {
+      "epoch": 3.9200998751560547,
+      "grad_norm": 0.271484375,
+      "learning_rate": 4.3223544457985735e-06,
+      "loss": 0.2315,
+      "step": 9420
+    },
+    {
+      "epoch": 3.9284228048272993,
+      "grad_norm": 0.2890625,
+      "learning_rate": 4.316751042412824e-06,
+      "loss": 0.2362,
+      "step": 9440
+    },
+    {
+      "epoch": 3.9367457344985435,
+      "grad_norm": 0.271484375,
+      "learning_rate": 4.311128231631745e-06,
+      "loss": 0.2343,
+      "step": 9460
+    },
+    {
+      "epoch": 3.9450686641697876,
+      "grad_norm": 0.328125,
+      "learning_rate": 4.305486073520865e-06,
+      "loss": 0.2349,
+      "step": 9480
+    },
+    {
+      "epoch": 3.9533915938410322,
+      "grad_norm": 0.275390625,
+      "learning_rate": 4.299824628352387e-06,
+      "loss": 0.2304,
+      "step": 9500
+    },
+    {
+      "epoch": 3.9533915938410322,
+      "eval_main_loss": 0.2382751852273941,
+      "eval_main_runtime": 6.3262,
+      "eval_main_samples_per_second": 30.034,
+      "eval_main_steps_per_second": 3.794,
+      "step": 9500
+    },
+    {
+      "epoch": 3.9533915938410322,
+      "eval_anatomy_loss": 2.832981586456299,
+      "eval_anatomy_runtime": 0.266,
+      "eval_anatomy_samples_per_second": 7.518,
+      "eval_anatomy_steps_per_second": 3.759,
+      "step": 9500
+    },
+    {
+      "epoch": 3.9533915938410322,
+      "eval_college_mathematics_loss": 2.0524649620056152,
+      "eval_college_mathematics_runtime": 0.2672,
+      "eval_college_mathematics_samples_per_second": 7.484,
+      "eval_college_mathematics_steps_per_second": 3.742,
+      "step": 9500
+    },
+    {
+      "epoch": 3.9533915938410322,
+      "eval_international_law_loss": 3.0676116943359375,
+      "eval_international_law_runtime": 0.2658,
+      "eval_international_law_samples_per_second": 7.525,
+      "eval_international_law_steps_per_second": 3.763,
+      "step": 9500
+    },
+    {
+      "epoch": 3.9617145235122764,
+      "grad_norm": 0.3125,
+      "learning_rate": 4.2941439566045536e-06,
+      "loss": 0.2308,
+      "step": 9520
+    },
+    {
+      "epoch": 3.9700374531835205,
+      "grad_norm": 0.291015625,
+      "learning_rate": 4.2884441189609915e-06,
+      "loss": 0.2317,
+      "step": 9540
+    },
+    {
+      "epoch": 3.978360382854765,
+      "grad_norm": 0.37109375,
+      "learning_rate": 4.282725176310065e-06,
+      "loss": 0.2318,
+      "step": 9560
+    },
+    {
+      "epoch": 3.9866833125260093,
+      "grad_norm": 0.26171875,
+      "learning_rate": 4.27698718974423e-06,
+      "loss": 0.2343,
+      "step": 9580
+    },
+    {
+      "epoch": 3.9950062421972534,
+      "grad_norm": 0.267578125,
+      "learning_rate": 4.271230220559378e-06,
+      "loss": 0.2294,
+      "step": 9600
+    },
+    {
+      "epoch": 3.9950062421972534,
+      "eval_main_loss": 0.23832967877388,
+      "eval_main_runtime": 6.3277,
+      "eval_main_samples_per_second": 30.027,
+      "eval_main_steps_per_second": 3.793,
+      "step": 9600
+    },
+    {
+      "epoch": 3.9950062421972534,
+      "eval_anatomy_loss": 2.8348305225372314,
+      "eval_anatomy_runtime": 0.2667,
+      "eval_anatomy_samples_per_second": 7.5,
+      "eval_anatomy_steps_per_second": 3.75,
+      "step": 9600
+    },
+    {
+      "epoch": 3.9950062421972534,
+      "eval_college_mathematics_loss": 2.0557780265808105,
+      "eval_college_mathematics_runtime": 0.2658,
+      "eval_college_mathematics_samples_per_second": 7.525,
+      "eval_college_mathematics_steps_per_second": 3.762,
+      "step": 9600
+    },
+    {
+      "epoch": 3.9950062421972534,
+      "eval_international_law_loss": 3.0656328201293945,
+      "eval_international_law_runtime": 0.2656,
+      "eval_international_law_samples_per_second": 7.53,
+      "eval_international_law_steps_per_second": 3.765,
+      "step": 9600
+    },
+    {
+      "epoch": 4.003329171868498,
+      "grad_norm": 0.27734375,
+      "learning_rate": 4.2654543302541796e-06,
+      "loss": 0.2352,
+      "step": 9620
+    },
+    {
+      "epoch": 4.011652101539742,
+      "grad_norm": 0.296875,
+      "learning_rate": 4.259659580529432e-06,
+      "loss": 0.2334,
+      "step": 9640
+    },
+    {
+      "epoch": 4.019975031210986,
+      "grad_norm": 0.2412109375,
+      "learning_rate": 4.253846033287398e-06,
+      "loss": 0.2308,
+      "step": 9660
+    },
+    {
+      "epoch": 4.028297960882231,
+      "grad_norm": 0.314453125,
+      "learning_rate": 4.248013750631143e-06,
+      "loss": 0.2353,
+      "step": 9680
+    },
+    {
+      "epoch": 4.036620890553475,
+      "grad_norm": 0.26953125,
+      "learning_rate": 4.242162794863872e-06,
+      "loss": 0.2312,
+      "step": 9700
+    },
+    {
+      "epoch": 4.036620890553475,
+      "eval_main_loss": 0.2384222000837326,
+      "eval_main_runtime": 6.3303,
+      "eval_main_samples_per_second": 30.014,
+      "eval_main_steps_per_second": 3.791,
+      "step": 9700
+    },
+    {
+      "epoch": 4.036620890553475,
+      "eval_anatomy_loss": 2.832493305206299,
+      "eval_anatomy_runtime": 0.2658,
+      "eval_anatomy_samples_per_second": 7.525,
+      "eval_anatomy_steps_per_second": 3.763,
+      "step": 9700
+    },
+    {
+      "epoch": 4.036620890553475,
+      "eval_college_mathematics_loss": 2.0524394512176514,
+      "eval_college_mathematics_runtime": 0.2664,
+      "eval_college_mathematics_samples_per_second": 7.508,
+      "eval_college_mathematics_steps_per_second": 3.754,
+      "step": 9700
+    },
+    {
+      "epoch": 4.036620890553475,
+      "eval_international_law_loss": 3.0663318634033203,
+      "eval_international_law_runtime": 0.2658,
+      "eval_international_law_samples_per_second": 7.525,
+      "eval_international_law_steps_per_second": 3.762,
+      "step": 9700
+    },
+    {
+      "epoch": 4.044943820224719,
+      "grad_norm": 0.298828125,
+      "learning_rate": 4.236293228488267e-06,
+      "loss": 0.2306,
+      "step": 9720
+    },
+    {
+      "epoch": 4.053266749895964,
+      "grad_norm": 0.298828125,
+      "learning_rate": 4.23040511420582e-06,
+      "loss": 0.2331,
+      "step": 9740
+    },
+    {
+      "epoch": 4.0615896795672075,
+      "grad_norm": 0.30859375,
+      "learning_rate": 4.224498514916152e-06,
+      "loss": 0.2312,
+      "step": 9760
+    },
+    {
+      "epoch": 4.069912609238452,
+      "grad_norm": 0.287109375,
+      "learning_rate": 4.218573493716359e-06,
+      "loss": 0.2321,
+      "step": 9780
+    },
+    {
+      "epoch": 4.078235538909696,
+      "grad_norm": 0.30078125,
+      "learning_rate": 4.212630113900322e-06,
+      "loss": 0.2333,
+      "step": 9800
+    },
+    {
+      "epoch": 4.078235538909696,
+      "eval_main_loss": 0.23829525709152222,
+      "eval_main_runtime": 6.3313,
+      "eval_main_samples_per_second": 30.01,
+      "eval_main_steps_per_second": 3.791,
+      "step": 9800
+    },
+    {
+      "epoch": 4.078235538909696,
+      "eval_anatomy_loss": 2.8339943885803223,
+      "eval_anatomy_runtime": 0.2661,
+      "eval_anatomy_samples_per_second": 7.515,
+      "eval_anatomy_steps_per_second": 3.757,
+      "step": 9800
+    },
+    {
+      "epoch": 4.078235538909696,
+      "eval_college_mathematics_loss": 2.0519936084747314,
+      "eval_college_mathematics_runtime": 0.2656,
+      "eval_college_mathematics_samples_per_second": 7.53,
+      "eval_college_mathematics_steps_per_second": 3.765,
+      "step": 9800
+    },
+    {
+      "epoch": 4.078235538909696,
+      "eval_international_law_loss": 3.06636905670166,
+      "eval_international_law_runtime": 0.2669,
+      "eval_international_law_samples_per_second": 7.494,
+      "eval_international_law_steps_per_second": 3.747,
+      "step": 9800
+    },
+    {
+      "epoch": 4.08655846858094,
+      "grad_norm": 0.287109375,
+      "learning_rate": 4.206668438958042e-06,
+      "loss": 0.2325,
+      "step": 9820
+    },
+    {
+      "epoch": 4.094881398252185,
+      "grad_norm": 0.294921875,
+      "learning_rate": 4.200688532574952e-06,
+      "loss": 0.2294,
+      "step": 9840
+    },
+    {
+      "epoch": 4.103204327923429,
+      "grad_norm": 0.275390625,
+      "learning_rate": 4.1946904586312485e-06,
+      "loss": 0.2326,
+      "step": 9860
+    },
+    {
+      "epoch": 4.111527257594673,
+      "grad_norm": 0.2578125,
+      "learning_rate": 4.188674281201198e-06,
+      "loss": 0.2348,
+      "step": 9880
+    },
+    {
+      "epoch": 4.119850187265918,
+      "grad_norm": 0.333984375,
+      "learning_rate": 4.182640064552456e-06,
+      "loss": 0.2324,
+      "step": 9900
+    },
+    {
+      "epoch": 4.119850187265918,
+      "eval_main_loss": 0.2383279949426651,
+      "eval_main_runtime": 6.3331,
+      "eval_main_samples_per_second": 30.001,
+      "eval_main_steps_per_second": 3.79,
+      "step": 9900
+    },
+    {
+      "epoch": 4.119850187265918,
+      "eval_anatomy_loss": 2.8347108364105225,
+      "eval_anatomy_runtime": 0.268,
+      "eval_anatomy_samples_per_second": 7.464,
+      "eval_anatomy_steps_per_second": 3.732,
+      "step": 9900
+    },
+    {
+      "epoch": 4.119850187265918,
+      "eval_college_mathematics_loss": 2.053701877593994,
+      "eval_college_mathematics_runtime": 0.266,
+      "eval_college_mathematics_samples_per_second": 7.519,
+      "eval_college_mathematics_steps_per_second": 3.759,
+      "step": 9900
+    },
+    {
+      "epoch": 4.119850187265918,
+      "eval_international_law_loss": 3.0677242279052734,
+      "eval_international_law_runtime": 0.2678,
+      "eval_international_law_samples_per_second": 7.469,
+      "eval_international_law_steps_per_second": 3.735,
+      "step": 9900
+    },
+    {
+      "epoch": 4.128173116937162,
+      "grad_norm": 0.244140625,
+      "learning_rate": 4.176587873145386e-06,
+      "loss": 0.2329,
+      "step": 9920
+    },
+    {
+      "epoch": 4.136496046608406,
+      "grad_norm": 0.3046875,
+      "learning_rate": 4.170517771632362e-06,
+      "loss": 0.2316,
+      "step": 9940
+    },
+    {
+      "epoch": 4.144818976279651,
+      "grad_norm": 0.2451171875,
+      "learning_rate": 4.164429824857086e-06,
+      "loss": 0.2336,
+      "step": 9960
+    },
+    {
+      "epoch": 4.1531419059508945,
+      "grad_norm": 0.255859375,
+      "learning_rate": 4.158324097853887e-06,
+      "loss": 0.2361,
+      "step": 9980
+    },
+    {
+      "epoch": 4.161464835622139,
+      "grad_norm": 0.26953125,
+      "learning_rate": 4.1522006558470365e-06,
+      "loss": 0.2359,
+      "step": 10000
+    },
+    {
+      "epoch": 4.161464835622139,
+      "eval_main_loss": 0.23838593065738678,
+      "eval_main_runtime": 6.3323,
+      "eval_main_samples_per_second": 30.005,
+      "eval_main_steps_per_second": 3.79,
+      "step": 10000
+    },
+    {
+      "epoch": 4.161464835622139,
+      "eval_anatomy_loss": 2.8330607414245605,
+      "eval_anatomy_runtime": 0.2671,
+      "eval_anatomy_samples_per_second": 7.488,
+      "eval_anatomy_steps_per_second": 3.744,
+      "step": 10000
+    },
+    {
+      "epoch": 4.161464835622139,
+      "eval_college_mathematics_loss": 2.04879093170166,
+      "eval_college_mathematics_runtime": 0.2649,
+      "eval_college_mathematics_samples_per_second": 7.551,
+      "eval_college_mathematics_steps_per_second": 3.776,
+      "step": 10000
+    },
+    {
+      "epoch": 4.161464835622139,
+      "eval_international_law_loss": 3.0646615028381348,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.506,
+      "eval_international_law_steps_per_second": 3.753,
+      "step": 10000
+    },
+    {
+      "epoch": 4.169787765293384,
+      "grad_norm": 0.322265625,
+      "learning_rate": 4.146059564250041e-06,
+      "loss": 0.2326,
+      "step": 10020
+    },
+    {
+      "epoch": 4.178110694964627,
+      "grad_norm": 0.2890625,
+      "learning_rate": 4.13990088866495e-06,
+      "loss": 0.2336,
+      "step": 10040
+    },
+    {
+      "epoch": 4.186433624635872,
+      "grad_norm": 0.310546875,
+      "learning_rate": 4.133724694881655e-06,
+      "loss": 0.2314,
+      "step": 10060
+    },
+    {
+      "epoch": 4.194756554307116,
+      "grad_norm": 0.306640625,
+      "learning_rate": 4.1275310488771855e-06,
+      "loss": 0.2286,
+      "step": 10080
+    },
+    {
+      "epoch": 4.20307948397836,
+      "grad_norm": 0.322265625,
+      "learning_rate": 4.1213200168149994e-06,
+      "loss": 0.2337,
+      "step": 10100
+    },
+    {
+      "epoch": 4.20307948397836,
+      "eval_main_loss": 0.23833875358104706,
+      "eval_main_runtime": 6.3313,
+      "eval_main_samples_per_second": 30.01,
+      "eval_main_steps_per_second": 3.791,
+      "step": 10100
+    },
+    {
+      "epoch": 4.20307948397836,
+      "eval_anatomy_loss": 2.8343758583068848,
+      "eval_anatomy_runtime": 0.2663,
+      "eval_anatomy_samples_per_second": 7.511,
+      "eval_anatomy_steps_per_second": 3.755,
+      "step": 10100
+    },
+    {
+      "epoch": 4.20307948397836,
+      "eval_college_mathematics_loss": 2.0506842136383057,
+      "eval_college_mathematics_runtime": 0.2649,
+      "eval_college_mathematics_samples_per_second": 7.551,
+      "eval_college_mathematics_steps_per_second": 3.775,
+      "step": 10100
+    },
+    {
+      "epoch": 4.20307948397836,
+      "eval_international_law_loss": 3.0656676292419434,
+      "eval_international_law_runtime": 0.267,
+      "eval_international_law_samples_per_second": 7.491,
+      "eval_international_law_steps_per_second": 3.745,
+      "step": 10100
+    },
+    {
+      "epoch": 4.211402413649605,
+      "grad_norm": 0.294921875,
+      "learning_rate": 4.115091665044284e-06,
+      "loss": 0.2332,
+      "step": 10120
+    },
+    {
+      "epoch": 4.219725343320849,
+      "grad_norm": 0.265625,
+      "learning_rate": 4.108846060099246e-06,
+      "loss": 0.2285,
+      "step": 10140
+    },
+    {
+      "epoch": 4.228048272992093,
+      "grad_norm": 0.3046875,
+      "learning_rate": 4.102583268698393e-06,
+      "loss": 0.2258,
+      "step": 10160
+    },
+    {
+      "epoch": 4.236371202663338,
+      "grad_norm": 0.328125,
+      "learning_rate": 4.096303357743834e-06,
+      "loss": 0.2307,
+      "step": 10180
+    },
+    {
+      "epoch": 4.2446941323345815,
+      "grad_norm": 0.2890625,
+      "learning_rate": 4.0900063943205485e-06,
+      "loss": 0.232,
+      "step": 10200
+    },
+    {
+      "epoch": 4.2446941323345815,
+      "eval_main_loss": 0.23833294212818146,
+      "eval_main_runtime": 6.3324,
+      "eval_main_samples_per_second": 30.004,
+      "eval_main_steps_per_second": 3.79,
+      "step": 10200
+    },
+    {
+      "epoch": 4.2446941323345815,
+      "eval_anatomy_loss": 2.8333046436309814,
+      "eval_anatomy_runtime": 0.2663,
+      "eval_anatomy_samples_per_second": 7.511,
+      "eval_anatomy_steps_per_second": 3.756,
+      "step": 10200
+    },
+    {
+      "epoch": 4.2446941323345815,
+      "eval_college_mathematics_loss": 2.0533056259155273,
+      "eval_college_mathematics_runtime": 0.2671,
+      "eval_college_mathematics_samples_per_second": 7.487,
+      "eval_college_mathematics_steps_per_second": 3.743,
+      "step": 10200
+    },
+    {
+      "epoch": 4.2446941323345815,
+      "eval_international_law_loss": 3.0654220581054688,
+      "eval_international_law_runtime": 0.2655,
+      "eval_international_law_samples_per_second": 7.533,
+      "eval_international_law_steps_per_second": 3.766,
+      "step": 10200
+    },
+    {
+      "epoch": 4.253017062005826,
+      "grad_norm": 0.263671875,
+      "learning_rate": 4.083692445695686e-06,
+      "loss": 0.2346,
+      "step": 10220
+    },
+    {
+      "epoch": 4.261339991677071,
+      "grad_norm": 0.2265625,
+      "learning_rate": 4.077361579317835e-06,
+      "loss": 0.2333,
+      "step": 10240
+    },
+    {
+      "epoch": 4.269662921348314,
+      "grad_norm": 0.2734375,
+      "learning_rate": 4.071013862816311e-06,
+      "loss": 0.2304,
+      "step": 10260
+    },
+    {
+      "epoch": 4.277985851019559,
+      "grad_norm": 0.2578125,
+      "learning_rate": 4.064649364000429e-06,
+      "loss": 0.235,
+      "step": 10280
+    },
+    {
+      "epoch": 4.286308780690804,
+      "grad_norm": 0.251953125,
+      "learning_rate": 4.058268150858779e-06,
+      "loss": 0.2325,
+      "step": 10300
+    },
+    {
+      "epoch": 4.286308780690804,
+      "eval_main_loss": 0.23827138543128967,
+      "eval_main_runtime": 6.325,
+      "eval_main_samples_per_second": 30.039,
+      "eval_main_steps_per_second": 3.794,
+      "step": 10300
+    },
+    {
+      "epoch": 4.286308780690804,
+      "eval_anatomy_loss": 2.832484006881714,
+      "eval_anatomy_runtime": 0.2664,
+      "eval_anatomy_samples_per_second": 7.509,
+      "eval_anatomy_steps_per_second": 3.754,
+      "step": 10300
+    },
+    {
+      "epoch": 4.286308780690804,
+      "eval_college_mathematics_loss": 2.050703287124634,
+      "eval_college_mathematics_runtime": 0.2662,
+      "eval_college_mathematics_samples_per_second": 7.512,
+      "eval_college_mathematics_steps_per_second": 3.756,
+      "step": 10300
+    },
+    {
+      "epoch": 4.286308780690804,
+      "eval_international_law_loss": 3.06673002243042,
+      "eval_international_law_runtime": 0.2654,
+      "eval_international_law_samples_per_second": 7.536,
+      "eval_international_law_steps_per_second": 3.768,
+      "step": 10300
+    },
+    {
+      "epoch": 4.294631710362047,
+      "grad_norm": 0.25390625,
+      "learning_rate": 4.051870291558505e-06,
+      "loss": 0.2288,
+      "step": 10320
+    },
+    {
+      "epoch": 4.302954640033292,
+      "grad_norm": 0.28125,
+      "learning_rate": 4.045455854444569e-06,
+      "loss": 0.232,
+      "step": 10340
+    },
+    {
+      "epoch": 4.3112775697045365,
+      "grad_norm": 0.255859375,
+      "learning_rate": 4.039024908039029e-06,
+      "loss": 0.2337,
+      "step": 10360
+    },
+    {
+      "epoch": 4.31960049937578,
+      "grad_norm": 0.3125,
+      "learning_rate": 4.0325775210402995e-06,
+      "loss": 0.2343,
+      "step": 10380
+    },
+    {
+      "epoch": 4.327923429047025,
+      "grad_norm": 0.267578125,
+      "learning_rate": 4.026113762322423e-06,
+      "loss": 0.2309,
+      "step": 10400
+    },
+    {
+      "epoch": 4.327923429047025,
+      "eval_main_loss": 0.23835839331150055,
+      "eval_main_runtime": 6.3532,
+      "eval_main_samples_per_second": 29.906,
+      "eval_main_steps_per_second": 3.778,
+      "step": 10400
+    },
+    {
+      "epoch": 4.327923429047025,
+      "eval_anatomy_loss": 2.8327348232269287,
+      "eval_anatomy_runtime": 0.2664,
+      "eval_anatomy_samples_per_second": 7.508,
+      "eval_anatomy_steps_per_second": 3.754,
+      "step": 10400
+    },
+    {
+      "epoch": 4.327923429047025,
+      "eval_college_mathematics_loss": 2.053480625152588,
+      "eval_college_mathematics_runtime": 0.2671,
+      "eval_college_mathematics_samples_per_second": 7.487,
+      "eval_college_mathematics_steps_per_second": 3.744,
+      "step": 10400
+    },
+    {
+      "epoch": 4.327923429047025,
+      "eval_international_law_loss": 3.0650157928466797,
+      "eval_international_law_runtime": 0.2677,
+      "eval_international_law_samples_per_second": 7.472,
+      "eval_international_law_steps_per_second": 3.736,
+      "step": 10400
+    },
+    {
+      "epoch": 4.3362463587182685,
+      "grad_norm": 0.26953125,
+      "learning_rate": 4.019633700934334e-06,
+      "loss": 0.2329,
+      "step": 10420
+    },
+    {
+      "epoch": 4.344569288389513,
+      "grad_norm": 0.302734375,
+      "learning_rate": 4.013137406099117e-06,
+      "loss": 0.2306,
+      "step": 10440
+    },
+    {
+      "epoch": 4.352892218060758,
+      "grad_norm": 0.271484375,
+      "learning_rate": 4.006624947213272e-06,
+      "loss": 0.2314,
+      "step": 10460
+    },
+    {
+      "epoch": 4.361215147732001,
+      "grad_norm": 0.259765625,
+      "learning_rate": 4.000096393845968e-06,
+      "loss": 0.2319,
+      "step": 10480
+    },
+    {
+      "epoch": 4.369538077403246,
+      "grad_norm": 0.271484375,
+      "learning_rate": 3.993551815738307e-06,
+      "loss": 0.2279,
+      "step": 10500
+    },
+    {
+      "epoch": 4.369538077403246,
+      "eval_main_loss": 0.23829132318496704,
+      "eval_main_runtime": 6.3595,
+      "eval_main_samples_per_second": 29.876,
+      "eval_main_steps_per_second": 3.774,
+      "step": 10500
+    },
+    {
+      "epoch": 4.369538077403246,
+      "eval_anatomy_loss": 2.8348493576049805,
+      "eval_anatomy_runtime": 0.2666,
+      "eval_anatomy_samples_per_second": 7.502,
+      "eval_anatomy_steps_per_second": 3.751,
+      "step": 10500
+    },
+    {
+      "epoch": 4.369538077403246,
+      "eval_college_mathematics_loss": 2.051111936569214,
+      "eval_college_mathematics_runtime": 0.2673,
+      "eval_college_mathematics_samples_per_second": 7.481,
+      "eval_college_mathematics_steps_per_second": 3.741,
+      "step": 10500
+    },
+    {
+      "epoch": 4.369538077403246,
+      "eval_international_law_loss": 3.0652830600738525,
+      "eval_international_law_runtime": 0.2673,
+      "eval_international_law_samples_per_second": 7.483,
+      "eval_international_law_steps_per_second": 3.742,
+      "step": 10500
+    },
+    {
+      "epoch": 4.3778610070744906,
+      "grad_norm": 0.267578125,
+      "learning_rate": 3.9869912828025735e-06,
+      "loss": 0.2286,
+      "step": 10520
+    },
+    {
+      "epoch": 4.386183936745734,
+      "grad_norm": 0.3203125,
+      "learning_rate": 3.980414865121486e-06,
+      "loss": 0.2314,
+      "step": 10540
+    },
+    {
+      "epoch": 4.394506866416979,
+      "grad_norm": 0.2578125,
+      "learning_rate": 3.973822632947455e-06,
+      "loss": 0.2295,
+      "step": 10560
+    },
+    {
+      "epoch": 4.4028297960882234,
+      "grad_norm": 0.337890625,
+      "learning_rate": 3.9672146567018275e-06,
+      "loss": 0.2325,
+      "step": 10580
+    },
+    {
+      "epoch": 4.411152725759467,
+      "grad_norm": 0.279296875,
+      "learning_rate": 3.9605910069741375e-06,
+      "loss": 0.2316,
+      "step": 10600
+    },
+    {
+      "epoch": 4.411152725759467,
+      "eval_main_loss": 0.23840609192848206,
+      "eval_main_runtime": 6.3547,
+      "eval_main_samples_per_second": 29.899,
+      "eval_main_steps_per_second": 3.777,
+      "step": 10600
+    },
+    {
+      "epoch": 4.411152725759467,
+      "eval_anatomy_loss": 2.8325912952423096,
+      "eval_anatomy_runtime": 0.2688,
+      "eval_anatomy_samples_per_second": 7.442,
+      "eval_anatomy_steps_per_second": 3.721,
+      "step": 10600
+    },
+    {
+      "epoch": 4.411152725759467,
+      "eval_college_mathematics_loss": 2.0505118370056152,
+      "eval_college_mathematics_runtime": 0.2654,
+      "eval_college_mathematics_samples_per_second": 7.537,
+      "eval_college_mathematics_steps_per_second": 3.769,
+      "step": 10600
+    },
+    {
+      "epoch": 4.411152725759467,
+      "eval_international_law_loss": 3.0657410621643066,
+      "eval_international_law_runtime": 0.2676,
+      "eval_international_law_samples_per_second": 7.475,
+      "eval_international_law_steps_per_second": 3.738,
+      "step": 10600
+    },
+    {
+      "epoch": 4.419475655430712,
+      "grad_norm": 0.287109375,
+      "learning_rate": 3.953951754521348e-06,
+      "loss": 0.2281,
+      "step": 10620
+    },
+    {
+      "epoch": 4.4277985851019555,
+      "grad_norm": 0.3046875,
+      "learning_rate": 3.947296970267098e-06,
+      "loss": 0.231,
+      "step": 10640
+    },
+    {
+      "epoch": 4.4361215147732,
+      "grad_norm": 0.2197265625,
+      "learning_rate": 3.940626725300949e-06,
+      "loss": 0.2298,
+      "step": 10660
+    },
+    {
+      "epoch": 4.444444444444445,
+      "grad_norm": 0.26953125,
+      "learning_rate": 3.933941090877615e-06,
+      "loss": 0.2314,
+      "step": 10680
+    },
+    {
+      "epoch": 4.452767374115688,
+      "grad_norm": 0.287109375,
+      "learning_rate": 3.927240138416212e-06,
+      "loss": 0.234,
+      "step": 10700
+    },
+    {
+      "epoch": 4.452767374115688,
+      "eval_main_loss": 0.2382432222366333,
+      "eval_main_runtime": 6.321,
+      "eval_main_samples_per_second": 30.059,
+      "eval_main_steps_per_second": 3.797,
+      "step": 10700
+    },
+    {
+      "epoch": 4.452767374115688,
+      "eval_anatomy_loss": 2.833085536956787,
+      "eval_anatomy_runtime": 0.2641,
+      "eval_anatomy_samples_per_second": 7.574,
+      "eval_anatomy_steps_per_second": 3.787,
+      "step": 10700
+    },
+    {
+      "epoch": 4.452767374115688,
+      "eval_college_mathematics_loss": 2.0491299629211426,
+      "eval_college_mathematics_runtime": 0.2669,
+      "eval_college_mathematics_samples_per_second": 7.494,
+      "eval_college_mathematics_steps_per_second": 3.747,
+      "step": 10700
+    },
+    {
+      "epoch": 4.452767374115688,
+      "eval_international_law_loss": 3.064594268798828,
+      "eval_international_law_runtime": 0.267,
+      "eval_international_law_samples_per_second": 7.492,
+      "eval_international_law_steps_per_second": 3.746,
+      "step": 10700
+    },
+    {
+      "epoch": 4.461090303786933,
+      "grad_norm": 0.2734375,
+      "learning_rate": 3.920523939499487e-06,
+      "loss": 0.2293,
+      "step": 10720
+    },
+    {
+      "epoch": 4.4694132334581775,
+      "grad_norm": 0.28125,
+      "learning_rate": 3.913792565873061e-06,
+      "loss": 0.2285,
+      "step": 10740
+    },
+    {
+      "epoch": 4.477736163129421,
+      "grad_norm": 0.275390625,
+      "learning_rate": 3.907046089444654e-06,
+      "loss": 0.2337,
+      "step": 10760
+    },
+    {
+      "epoch": 4.486059092800666,
+      "grad_norm": 0.34375,
+      "learning_rate": 3.900284582283323e-06,
+      "loss": 0.2325,
+      "step": 10780
+    },
+    {
+      "epoch": 4.49438202247191,
+      "grad_norm": 0.369140625,
+      "learning_rate": 3.8935081166186935e-06,
+      "loss": 0.2343,
+      "step": 10800
+    },
+    {
+      "epoch": 4.49438202247191,
+      "eval_main_loss": 0.23841159045696259,
+      "eval_main_runtime": 6.3194,
+      "eval_main_samples_per_second": 30.066,
+      "eval_main_steps_per_second": 3.798,
+      "step": 10800
+    },
+    {
+      "epoch": 4.49438202247191,
+      "eval_anatomy_loss": 2.8321523666381836,
+      "eval_anatomy_runtime": 0.2659,
+      "eval_anatomy_samples_per_second": 7.521,
+      "eval_anatomy_steps_per_second": 3.76,
+      "step": 10800
+    },
+    {
+      "epoch": 4.49438202247191,
+      "eval_college_mathematics_loss": 2.050726890563965,
+      "eval_college_mathematics_runtime": 0.2654,
+      "eval_college_mathematics_samples_per_second": 7.536,
+      "eval_college_mathematics_steps_per_second": 3.768,
+      "step": 10800
+    },
+    {
+      "epoch": 4.49438202247191,
+      "eval_international_law_loss": 3.0631279945373535,
+      "eval_international_law_runtime": 0.2661,
+      "eval_international_law_samples_per_second": 7.517,
+      "eval_international_law_steps_per_second": 3.759,
+      "step": 10800
+    },
+    {
+      "epoch": 4.502704952143154,
+      "grad_norm": 0.314453125,
+      "learning_rate": 3.88671676484018e-06,
+      "loss": 0.2325,
+      "step": 10820
+    },
+    {
+      "epoch": 4.511027881814399,
+      "grad_norm": 0.322265625,
+      "learning_rate": 3.87991059949622e-06,
+      "loss": 0.2315,
+      "step": 10840
+    },
+    {
+      "epoch": 4.519350811485643,
+      "grad_norm": 0.29296875,
+      "learning_rate": 3.873089693293497e-06,
+      "loss": 0.2325,
+      "step": 10860
+    },
+    {
+      "epoch": 4.527673741156887,
+      "grad_norm": 0.25,
+      "learning_rate": 3.866254119096161e-06,
+      "loss": 0.2298,
+      "step": 10880
+    },
+    {
+      "epoch": 4.535996670828132,
+      "grad_norm": 0.275390625,
+      "learning_rate": 3.8594039499250545e-06,
+      "loss": 0.232,
+      "step": 10900
+    },
+    {
+      "epoch": 4.535996670828132,
+      "eval_main_loss": 0.2383212000131607,
+      "eval_main_runtime": 6.3117,
+      "eval_main_samples_per_second": 30.103,
+      "eval_main_steps_per_second": 3.802,
+      "step": 10900
+    },
+    {
+      "epoch": 4.535996670828132,
+      "eval_anatomy_loss": 2.833933115005493,
+      "eval_anatomy_runtime": 0.2655,
+      "eval_anatomy_samples_per_second": 7.534,
+      "eval_anatomy_steps_per_second": 3.767,
+      "step": 10900
+    },
+    {
+      "epoch": 4.535996670828132,
+      "eval_college_mathematics_loss": 2.0495407581329346,
+      "eval_college_mathematics_runtime": 0.2654,
+      "eval_college_mathematics_samples_per_second": 7.537,
+      "eval_college_mathematics_steps_per_second": 3.768,
+      "step": 10900
+    },
+    {
+      "epoch": 4.535996670828132,
+      "eval_international_law_loss": 3.0634400844573975,
+      "eval_international_law_runtime": 0.2654,
+      "eval_international_law_samples_per_second": 7.537,
+      "eval_international_law_steps_per_second": 3.768,
+      "step": 10900
+    },
+    {
+      "epoch": 4.544319600499376,
+      "grad_norm": 0.248046875,
+      "learning_rate": 3.852539258956931e-06,
+      "loss": 0.2345,
+      "step": 10920
+    },
+    {
+      "epoch": 4.55264253017062,
+      "grad_norm": 0.302734375,
+      "learning_rate": 3.845660119523671e-06,
+      "loss": 0.2315,
+      "step": 10940
+    },
+    {
+      "epoch": 4.5609654598418645,
+      "grad_norm": 0.27734375,
+      "learning_rate": 3.8387666051114995e-06,
+      "loss": 0.232,
+      "step": 10960
+    },
+    {
+      "epoch": 4.569288389513108,
+      "grad_norm": 0.30078125,
+      "learning_rate": 3.831858789360206e-06,
+      "loss": 0.2332,
+      "step": 10980
+    },
+    {
+      "epoch": 4.577611319184353,
+      "grad_norm": 0.26171875,
+      "learning_rate": 3.824936746062349e-06,
+      "loss": 0.2349,
+      "step": 11000
+    },
+    {
+      "epoch": 4.577611319184353,
+      "eval_main_loss": 0.23827816545963287,
+      "eval_main_runtime": 6.3591,
+      "eval_main_samples_per_second": 29.878,
+      "eval_main_steps_per_second": 3.774,
+      "step": 11000
+    },
+    {
+      "epoch": 4.577611319184353,
+      "eval_anatomy_loss": 2.83370304107666,
+      "eval_anatomy_runtime": 0.2663,
+      "eval_anatomy_samples_per_second": 7.51,
+      "eval_anatomy_steps_per_second": 3.755,
+      "step": 11000
+    },
+    {
+      "epoch": 4.577611319184353,
+      "eval_college_mathematics_loss": 2.051502227783203,
+      "eval_college_mathematics_runtime": 0.2674,
+      "eval_college_mathematics_samples_per_second": 7.478,
+      "eval_college_mathematics_steps_per_second": 3.739,
+      "step": 11000
+    },
+    {
+      "epoch": 4.577611319184353,
+      "eval_international_law_loss": 3.0672590732574463,
+      "eval_international_law_runtime": 0.268,
+      "eval_international_law_samples_per_second": 7.463,
+      "eval_international_law_steps_per_second": 3.731,
+      "step": 11000
+    },
+    {
+      "epoch": 4.585934248855597,
+      "grad_norm": 0.291015625,
+      "learning_rate": 3.818000549162474e-06,
+      "loss": 0.2318,
+      "step": 11020
+    },
+    {
+      "epoch": 4.594257178526841,
+      "grad_norm": 0.30078125,
+      "learning_rate": 3.811050272756324e-06,
+      "loss": 0.2348,
+      "step": 11040
+    },
+    {
+      "epoch": 4.602580108198086,
+      "grad_norm": 0.28515625,
+      "learning_rate": 3.804085991090044e-06,
+      "loss": 0.2342,
+      "step": 11060
+    },
+    {
+      "epoch": 4.61090303786933,
+      "grad_norm": 0.3125,
+      "learning_rate": 3.797107778559389e-06,
+      "loss": 0.2307,
+      "step": 11080
+    },
+    {
+      "epoch": 4.619225967540574,
+      "grad_norm": 0.2451171875,
+      "learning_rate": 3.7901157097089315e-06,
+      "loss": 0.228,
+      "step": 11100
+    },
+    {
+      "epoch": 4.619225967540574,
+      "eval_main_loss": 0.23824092745780945,
+      "eval_main_runtime": 6.3522,
+      "eval_main_samples_per_second": 29.911,
+      "eval_main_steps_per_second": 3.778,
+      "step": 11100
+    },
+    {
+      "epoch": 4.619225967540574,
+      "eval_anatomy_loss": 2.834642171859741,
+      "eval_anatomy_runtime": 0.2679,
+      "eval_anatomy_samples_per_second": 7.465,
+      "eval_anatomy_steps_per_second": 3.732,
+      "step": 11100
+    },
+    {
+      "epoch": 4.619225967540574,
+      "eval_college_mathematics_loss": 2.054388999938965,
+      "eval_college_mathematics_runtime": 0.2658,
+      "eval_college_mathematics_samples_per_second": 7.525,
+      "eval_college_mathematics_steps_per_second": 3.763,
+      "step": 11100
+    },
+    {
+      "epoch": 4.619225967540574,
+      "eval_international_law_loss": 3.0648648738861084,
+      "eval_international_law_runtime": 0.2679,
+      "eval_international_law_samples_per_second": 7.465,
+      "eval_international_law_steps_per_second": 3.732,
+      "step": 11100
+    },
+    {
+      "epoch": 4.627548897211819,
+      "grad_norm": 0.275390625,
+      "learning_rate": 3.7831098592312643e-06,
+      "loss": 0.2368,
+      "step": 11120
+    },
+    {
+      "epoch": 4.635871826883063,
+      "grad_norm": 0.26171875,
+      "learning_rate": 3.7760903019662008e-06,
+      "loss": 0.2292,
+      "step": 11140
+    },
+    {
+      "epoch": 4.644194756554307,
+      "grad_norm": 0.298828125,
+      "learning_rate": 3.7690571128999775e-06,
+      "loss": 0.2371,
+      "step": 11160
+    },
+    {
+      "epoch": 4.6525176862255515,
+      "grad_norm": 0.2734375,
+      "learning_rate": 3.7620103671644516e-06,
+      "loss": 0.2305,
+      "step": 11180
+    },
+    {
+      "epoch": 4.660840615896795,
+      "grad_norm": 0.26953125,
+      "learning_rate": 3.7549501400362996e-06,
+      "loss": 0.2265,
+      "step": 11200
+    },
+    {
+      "epoch": 4.660840615896795,
+      "eval_main_loss": 0.2383652925491333,
+      "eval_main_runtime": 6.3394,
+      "eval_main_samples_per_second": 29.971,
+      "eval_main_steps_per_second": 3.786,
+      "step": 11200
+    },
+    {
+      "epoch": 4.660840615896795,
+      "eval_anatomy_loss": 2.8311378955841064,
+      "eval_anatomy_runtime": 0.2675,
+      "eval_anatomy_samples_per_second": 7.477,
+      "eval_anatomy_steps_per_second": 3.738,
+      "step": 11200
+    },
+    {
+      "epoch": 4.660840615896795,
+      "eval_college_mathematics_loss": 2.052487373352051,
+      "eval_college_mathematics_runtime": 0.267,
+      "eval_college_mathematics_samples_per_second": 7.492,
+      "eval_college_mathematics_steps_per_second": 3.746,
+      "step": 11200
+    },
+    {
+      "epoch": 4.660840615896795,
+      "eval_international_law_loss": 3.064990758895874,
+      "eval_international_law_runtime": 0.2666,
+      "eval_international_law_samples_per_second": 7.501,
+      "eval_international_law_steps_per_second": 3.751,
+      "step": 11200
+    },
+    {
+      "epoch": 4.66916354556804,
+      "grad_norm": 0.27734375,
+      "learning_rate": 3.7478765069362122e-06,
+      "loss": 0.2371,
+      "step": 11220
+    },
+    {
+      "epoch": 4.677486475239284,
+      "grad_norm": 0.32421875,
+      "learning_rate": 3.7407895434280893e-06,
+      "loss": 0.2293,
+      "step": 11240
+    },
+    {
+      "epoch": 4.685809404910529,
+      "grad_norm": 0.333984375,
+      "learning_rate": 3.7336893252182343e-06,
+      "loss": 0.2329,
+      "step": 11260
+    },
+    {
+      "epoch": 4.694132334581773,
+      "grad_norm": 0.28515625,
+      "learning_rate": 3.72657592815454e-06,
+      "loss": 0.2364,
+      "step": 11280
+    },
+    {
+      "epoch": 4.702455264253017,
+      "grad_norm": 0.259765625,
+      "learning_rate": 3.719449428225685e-06,
+      "loss": 0.2309,
+      "step": 11300
+    },
+    {
+      "epoch": 4.702455264253017,
+      "eval_main_loss": 0.23832421004772186,
+      "eval_main_runtime": 6.3244,
+      "eval_main_samples_per_second": 30.042,
+      "eval_main_steps_per_second": 3.795,
+      "step": 11300
+    },
+    {
+      "epoch": 4.702455264253017,
+      "eval_anatomy_loss": 2.8343684673309326,
+      "eval_anatomy_runtime": 0.2667,
+      "eval_anatomy_samples_per_second": 7.5,
+      "eval_anatomy_steps_per_second": 3.75,
+      "step": 11300
+    },
+    {
+      "epoch": 4.702455264253017,
+      "eval_college_mathematics_loss": 2.049694776535034,
+      "eval_college_mathematics_runtime": 0.2665,
+      "eval_college_mathematics_samples_per_second": 7.504,
+      "eval_college_mathematics_steps_per_second": 3.752,
+      "step": 11300
+    },
+    {
+      "epoch": 4.702455264253017,
+      "eval_international_law_loss": 3.0649161338806152,
+      "eval_international_law_runtime": 0.2654,
+      "eval_international_law_samples_per_second": 7.536,
+      "eval_international_law_steps_per_second": 3.768,
+      "step": 11300
+    },
+    {
+      "epoch": 4.710778193924261,
+      "grad_norm": 0.31640625,
+      "learning_rate": 3.712309901560316e-06,
+      "loss": 0.2332,
+      "step": 11320
+    },
+    {
+      "epoch": 4.719101123595506,
+      "grad_norm": 0.279296875,
+      "learning_rate": 3.7051574244262412e-06,
+      "loss": 0.2321,
+      "step": 11340
+    },
+    {
+      "epoch": 4.72742405326675,
+      "grad_norm": 0.22265625,
+      "learning_rate": 3.6979920732296085e-06,
+      "loss": 0.2368,
+      "step": 11360
+    },
+    {
+      "epoch": 4.735746982937994,
+      "grad_norm": 0.25,
+      "learning_rate": 3.690813924514095e-06,
+      "loss": 0.2292,
+      "step": 11380
+    },
+    {
+      "epoch": 4.7440699126092385,
+      "grad_norm": 0.283203125,
+      "learning_rate": 3.6836230549600853e-06,
+      "loss": 0.231,
+      "step": 11400
+    },
+    {
+      "epoch": 4.7440699126092385,
+      "eval_main_loss": 0.23822040855884552,
+      "eval_main_runtime": 6.3505,
+      "eval_main_samples_per_second": 29.919,
+      "eval_main_steps_per_second": 3.779,
+      "step": 11400
+    },
+    {
+      "epoch": 4.7440699126092385,
+      "eval_anatomy_loss": 2.8312675952911377,
+      "eval_anatomy_runtime": 0.2674,
+      "eval_anatomy_samples_per_second": 7.48,
+      "eval_anatomy_steps_per_second": 3.74,
+      "step": 11400
+    },
+    {
+      "epoch": 4.7440699126092385,
+      "eval_college_mathematics_loss": 2.0531413555145264,
+      "eval_college_mathematics_runtime": 0.2667,
+      "eval_college_mathematics_samples_per_second": 7.498,
+      "eval_college_mathematics_steps_per_second": 3.749,
+      "step": 11400
+    },
+    {
+      "epoch": 4.7440699126092385,
+      "eval_international_law_loss": 3.0663352012634277,
+      "eval_international_law_runtime": 0.2678,
+      "eval_international_law_samples_per_second": 7.467,
+      "eval_international_law_steps_per_second": 3.733,
+      "step": 11400
+    },
+    {
+      "epoch": 4.752392842280483,
+      "grad_norm": 0.259765625,
+      "learning_rate": 3.676419541383855e-06,
+      "loss": 0.2289,
+      "step": 11420
+    },
+    {
+      "epoch": 4.760715771951727,
+      "grad_norm": 0.2734375,
+      "learning_rate": 3.6692034607367486e-06,
+      "loss": 0.2305,
+      "step": 11440
+    },
+    {
+      "epoch": 4.769038701622971,
+      "grad_norm": 0.263671875,
+      "learning_rate": 3.6619748901043583e-06,
+      "loss": 0.2279,
+      "step": 11460
+    },
+    {
+      "epoch": 4.777361631294216,
+      "grad_norm": 0.30078125,
+      "learning_rate": 3.6547339067057007e-06,
+      "loss": 0.2382,
+      "step": 11480
+    },
+    {
+      "epoch": 4.78568456096546,
+      "grad_norm": 0.328125,
+      "learning_rate": 3.647480587892391e-06,
+      "loss": 0.2302,
+      "step": 11500
+    },
+    {
+      "epoch": 4.78568456096546,
+      "eval_main_loss": 0.23836463689804077,
+      "eval_main_runtime": 6.3523,
+      "eval_main_samples_per_second": 29.911,
+      "eval_main_steps_per_second": 3.778,
+      "step": 11500
+    },
+    {
+      "epoch": 4.78568456096546,
+      "eval_anatomy_loss": 2.8309946060180664,
+      "eval_anatomy_runtime": 0.2675,
+      "eval_anatomy_samples_per_second": 7.477,
+      "eval_anatomy_steps_per_second": 3.738,
+      "step": 11500
+    },
+    {
+      "epoch": 4.78568456096546,
+      "eval_college_mathematics_loss": 2.0509681701660156,
+      "eval_college_mathematics_runtime": 0.2662,
+      "eval_college_mathematics_samples_per_second": 7.513,
+      "eval_college_mathematics_steps_per_second": 3.757,
+      "step": 11500
+    },
+    {
+      "epoch": 4.78568456096546,
+      "eval_international_law_loss": 3.06599760055542,
+      "eval_international_law_runtime": 0.2671,
+      "eval_international_law_samples_per_second": 7.487,
+      "eval_international_law_steps_per_second": 3.743,
+      "step": 11500
+    },
+    {
+      "epoch": 4.794007490636704,
+      "grad_norm": 0.2470703125,
+      "learning_rate": 3.640215011147815e-06,
+      "loss": 0.2276,
+      "step": 11520
+    },
+    {
+      "epoch": 4.802330420307948,
+      "grad_norm": 0.2734375,
+      "learning_rate": 3.632937254086308e-06,
+      "loss": 0.2298,
+      "step": 11540
+    },
+    {
+      "epoch": 4.810653349979193,
+      "grad_norm": 0.28125,
+      "learning_rate": 3.6256473944523175e-06,
+      "loss": 0.2362,
+      "step": 11560
+    },
+    {
+      "epoch": 4.818976279650437,
+      "grad_norm": 0.255859375,
+      "learning_rate": 3.6183455101195785e-06,
+      "loss": 0.2301,
+      "step": 11580
+    },
+    {
+      "epoch": 4.827299209321681,
+      "grad_norm": 0.30078125,
+      "learning_rate": 3.611031679090278e-06,
+      "loss": 0.2313,
+      "step": 11600
+    },
+    {
+      "epoch": 4.827299209321681,
+      "eval_main_loss": 0.23831431567668915,
+      "eval_main_runtime": 6.3233,
+      "eval_main_samples_per_second": 30.048,
+      "eval_main_steps_per_second": 3.795,
+      "step": 11600
+    },
+    {
+      "epoch": 4.827299209321681,
+      "eval_anatomy_loss": 2.8330864906311035,
+      "eval_anatomy_runtime": 0.2671,
+      "eval_anatomy_samples_per_second": 7.489,
+      "eval_anatomy_steps_per_second": 3.744,
+      "step": 11600
+    },
+    {
+      "epoch": 4.827299209321681,
+      "eval_college_mathematics_loss": 2.0517561435699463,
+      "eval_college_mathematics_runtime": 0.2651,
+      "eval_college_mathematics_samples_per_second": 7.544,
+      "eval_college_mathematics_steps_per_second": 3.772,
+      "step": 11600
+    },
+    {
+      "epoch": 4.827299209321681,
+      "eval_international_law_loss": 3.066469669342041,
+      "eval_international_law_runtime": 0.2658,
+      "eval_international_law_samples_per_second": 7.523,
+      "eval_international_law_steps_per_second": 3.762,
+      "step": 11600
+    },
+    {
+      "epoch": 4.8356221389929255,
+      "grad_norm": 0.2890625,
+      "learning_rate": 3.603705979494225e-06,
+      "loss": 0.2306,
+      "step": 11620
+    },
+    {
+      "epoch": 4.84394506866417,
+      "grad_norm": 0.287109375,
+      "learning_rate": 3.5963684895880123e-06,
+      "loss": 0.229,
+      "step": 11640
+    },
+    {
+      "epoch": 4.852267998335414,
+      "grad_norm": 0.341796875,
+      "learning_rate": 3.589019287754183e-06,
+      "loss": 0.2244,
+      "step": 11660
+    },
+    {
+      "epoch": 4.860590928006658,
+      "grad_norm": 0.326171875,
+      "learning_rate": 3.581658452500394e-06,
+      "loss": 0.2329,
+      "step": 11680
+    },
+    {
+      "epoch": 4.868913857677903,
+      "grad_norm": 0.2421875,
+      "learning_rate": 3.574286062458574e-06,
+      "loss": 0.2329,
+      "step": 11700
+    },
+    {
+      "epoch": 4.868913857677903,
+      "eval_main_loss": 0.2383066862821579,
+      "eval_main_runtime": 6.3157,
+      "eval_main_samples_per_second": 30.084,
+      "eval_main_steps_per_second": 3.8,
+      "step": 11700
+    },
+    {
+      "epoch": 4.868913857677903,
+      "eval_anatomy_loss": 2.832721710205078,
+      "eval_anatomy_runtime": 0.2659,
+      "eval_anatomy_samples_per_second": 7.52,
+      "eval_anatomy_steps_per_second": 3.76,
+      "step": 11700
+    },
+    {
+      "epoch": 4.868913857677903,
+      "eval_college_mathematics_loss": 2.048736095428467,
+      "eval_college_mathematics_runtime": 0.2649,
+      "eval_college_mathematics_samples_per_second": 7.549,
+      "eval_college_mathematics_steps_per_second": 3.774,
+      "step": 11700
+    },
+    {
+      "epoch": 4.868913857677903,
+      "eval_international_law_loss": 3.0647518634796143,
+      "eval_international_law_runtime": 0.2646,
+      "eval_international_law_samples_per_second": 7.558,
+      "eval_international_law_steps_per_second": 3.779,
+      "step": 11700
+    },
+    {
+      "epoch": 4.877236787349147,
+      "grad_norm": 0.259765625,
+      "learning_rate": 3.5669021963840863e-06,
+      "loss": 0.2342,
+      "step": 11720
+    },
+    {
+      "epoch": 4.885559717020391,
+      "grad_norm": 0.26171875,
+      "learning_rate": 3.559506933154886e-06,
+      "loss": 0.2326,
+      "step": 11740
+    },
+    {
+      "epoch": 4.893882646691635,
+      "grad_norm": 0.28515625,
+      "learning_rate": 3.552100351770679e-06,
+      "loss": 0.2283,
+      "step": 11760
+    },
+    {
+      "epoch": 4.90220557636288,
+      "grad_norm": 0.265625,
+      "learning_rate": 3.544682531352076e-06,
+      "loss": 0.2359,
+      "step": 11780
+    },
+    {
+      "epoch": 4.910528506034124,
+      "grad_norm": 0.314453125,
+      "learning_rate": 3.53725355113975e-06,
+      "loss": 0.2338,
+      "step": 11800
+    },
+    {
+      "epoch": 4.910528506034124,
+      "eval_main_loss": 0.23833368718624115,
+      "eval_main_runtime": 6.3601,
+      "eval_main_samples_per_second": 29.874,
+      "eval_main_steps_per_second": 3.774,
+      "step": 11800
+    },
+    {
+      "epoch": 4.910528506034124,
+      "eval_anatomy_loss": 2.8294174671173096,
+      "eval_anatomy_runtime": 0.2674,
+      "eval_anatomy_samples_per_second": 7.479,
+      "eval_anatomy_steps_per_second": 3.74,
+      "step": 11800
+    },
+    {
+      "epoch": 4.910528506034124,
+      "eval_college_mathematics_loss": 2.054882764816284,
+      "eval_college_mathematics_runtime": 0.2668,
+      "eval_college_mathematics_samples_per_second": 7.495,
+      "eval_college_mathematics_steps_per_second": 3.747,
+      "step": 11800
+    },
+    {
+      "epoch": 4.910528506034124,
+      "eval_international_law_loss": 3.0666239261627197,
+      "eval_international_law_runtime": 0.2669,
+      "eval_international_law_samples_per_second": 7.493,
+      "eval_international_law_steps_per_second": 3.746,
+      "step": 11800
+    },
+    {
+      "epoch": 4.918851435705369,
+      "grad_norm": 0.2451171875,
+      "learning_rate": 3.529813490493586e-06,
+      "loss": 0.2312,
+      "step": 11820
+    },
+    {
+      "epoch": 4.9271743653766125,
+      "grad_norm": 0.298828125,
+      "learning_rate": 3.5223624288918368e-06,
+      "loss": 0.2272,
+      "step": 11840
+    },
+    {
+      "epoch": 4.935497295047857,
+      "grad_norm": 0.390625,
+      "learning_rate": 3.514900445930273e-06,
+      "loss": 0.2308,
+      "step": 11860
+    },
+    {
+      "epoch": 4.943820224719101,
+      "grad_norm": 0.306640625,
+      "learning_rate": 3.507427621321331e-06,
+      "loss": 0.2286,
+      "step": 11880
+    },
+    {
+      "epoch": 4.952143154390345,
+      "grad_norm": 0.3046875,
+      "learning_rate": 3.4999440348932644e-06,
+      "loss": 0.2363,
+      "step": 11900
+    },
+    {
+      "epoch": 4.952143154390345,
+      "eval_main_loss": 0.23830710351467133,
+      "eval_main_runtime": 6.3539,
+      "eval_main_samples_per_second": 29.903,
+      "eval_main_steps_per_second": 3.777,
+      "step": 11900
+    },
+    {
+      "epoch": 4.952143154390345,
+      "eval_anatomy_loss": 2.832731008529663,
+      "eval_anatomy_runtime": 0.2683,
+      "eval_anatomy_samples_per_second": 7.453,
+      "eval_anatomy_steps_per_second": 3.727,
+      "step": 11900
+    },
+    {
+      "epoch": 4.952143154390345,
+      "eval_college_mathematics_loss": 2.0541601181030273,
+      "eval_college_mathematics_runtime": 0.2669,
+      "eval_college_mathematics_samples_per_second": 7.492,
+      "eval_college_mathematics_steps_per_second": 3.746,
+      "step": 11900
+    },
+    {
+      "epoch": 4.952143154390345,
+      "eval_international_law_loss": 3.064253091812134,
+      "eval_international_law_runtime": 0.2674,
+      "eval_international_law_samples_per_second": 7.48,
+      "eval_international_law_steps_per_second": 3.74,
+      "step": 11900
+    },
+    {
+      "epoch": 4.96046608406159,
+      "grad_norm": 0.271484375,
+      "learning_rate": 3.4924497665892886e-06,
+      "loss": 0.2364,
+      "step": 11920
+    },
+    {
+      "epoch": 4.968789013732834,
+      "grad_norm": 0.271484375,
+      "learning_rate": 3.484944896466727e-06,
+      "loss": 0.2331,
+      "step": 11940
+    },
+    {
+      "epoch": 4.977111943404078,
+      "grad_norm": 0.2734375,
+      "learning_rate": 3.4774295046961593e-06,
+      "loss": 0.2344,
+      "step": 11960
+    },
+    {
+      "epoch": 4.985434873075323,
+      "grad_norm": 0.28515625,
+      "learning_rate": 3.4699036715605595e-06,
+      "loss": 0.2307,
+      "step": 11980
+    },
+    {
+      "epoch": 4.9937578027465666,
+      "grad_norm": 0.294921875,
+      "learning_rate": 3.4623674774544435e-06,
+      "loss": 0.2271,
+      "step": 12000
+    },
+    {
+      "epoch": 4.9937578027465666,
+      "eval_main_loss": 0.23841014504432678,
+      "eval_main_runtime": 6.3245,
+      "eval_main_samples_per_second": 30.042,
+      "eval_main_steps_per_second": 3.795,
+      "step": 12000
+    },
+    {
+      "epoch": 4.9937578027465666,
+      "eval_anatomy_loss": 2.8334553241729736,
+      "eval_anatomy_runtime": 0.2654,
+      "eval_anatomy_samples_per_second": 7.535,
+      "eval_anatomy_steps_per_second": 3.767,
+      "step": 12000
+    },
+    {
+      "epoch": 4.9937578027465666,
+      "eval_college_mathematics_loss": 2.050748348236084,
+      "eval_college_mathematics_runtime": 0.2648,
+      "eval_college_mathematics_samples_per_second": 7.553,
+      "eval_college_mathematics_steps_per_second": 3.776,
+      "step": 12000
+    },
+    {
+      "epoch": 4.9937578027465666,
+      "eval_international_law_loss": 3.066818952560425,
+      "eval_international_law_runtime": 0.2674,
+      "eval_international_law_samples_per_second": 7.479,
+      "eval_international_law_steps_per_second": 3.74,
+      "step": 12000
+    },
+    {
+      "epoch": 5.002080732417811,
+      "grad_norm": 0.267578125,
+      "learning_rate": 3.454821002883007e-06,
+      "loss": 0.2312,
+      "step": 12020
+    },
+    {
+      "epoch": 5.010403662089056,
+      "grad_norm": 0.29296875,
+      "learning_rate": 3.4472643284612656e-06,
+      "loss": 0.2312,
+      "step": 12040
+    },
+    {
+      "epoch": 5.0187265917602994,
+      "grad_norm": 0.287109375,
+      "learning_rate": 3.439697534913197e-06,
+      "loss": 0.2339,
+      "step": 12060
+    },
+    {
+      "epoch": 5.027049521431544,
+      "grad_norm": 0.23046875,
+      "learning_rate": 3.4321207030708725e-06,
+      "loss": 0.2329,
+      "step": 12080
+    },
+    {
+      "epoch": 5.035372451102788,
+      "grad_norm": 0.2734375,
+      "learning_rate": 3.4245339138736023e-06,
+      "loss": 0.2298,
+      "step": 12100
+    },
+    {
+      "epoch": 5.035372451102788,
+      "eval_main_loss": 0.23836223781108856,
+      "eval_main_runtime": 6.3231,
+      "eval_main_samples_per_second": 30.049,
+      "eval_main_steps_per_second": 3.796,
+      "step": 12100
+    },
+    {
+      "epoch": 5.035372451102788,
+      "eval_anatomy_loss": 2.8340742588043213,
+      "eval_anatomy_runtime": 0.2664,
+      "eval_anatomy_samples_per_second": 7.507,
+      "eval_anatomy_steps_per_second": 3.754,
+      "step": 12100
+    },
+    {
+      "epoch": 5.035372451102788,
+      "eval_college_mathematics_loss": 2.0523436069488525,
+      "eval_college_mathematics_runtime": 0.2651,
+      "eval_college_mathematics_samples_per_second": 7.545,
+      "eval_college_mathematics_steps_per_second": 3.773,
+      "step": 12100
+    },
+    {
+      "epoch": 5.035372451102788,
+      "eval_international_law_loss": 3.0669960975646973,
+      "eval_international_law_runtime": 0.2656,
+      "eval_international_law_samples_per_second": 7.529,
+      "eval_international_law_steps_per_second": 3.765,
+      "step": 12100
+    },
+    {
+      "epoch": 5.043695380774032,
+      "grad_norm": 0.255859375,
+      "learning_rate": 3.416937248367061e-06,
+      "loss": 0.2311,
+      "step": 12120
+    },
+    {
+      "epoch": 5.052018310445277,
+      "grad_norm": 0.25390625,
+      "learning_rate": 3.409330787702428e-06,
+      "loss": 0.2297,
+      "step": 12140
+    },
+    {
+      "epoch": 5.060341240116521,
+      "grad_norm": 0.263671875,
+      "learning_rate": 3.4017146131355205e-06,
+      "loss": 0.2329,
+      "step": 12160
+    },
+    {
+      "epoch": 5.068664169787765,
+      "grad_norm": 0.267578125,
+      "learning_rate": 3.394088806025925e-06,
+      "loss": 0.2279,
+      "step": 12180
+    },
+    {
+      "epoch": 5.07698709945901,
+      "grad_norm": 0.267578125,
+      "learning_rate": 3.3864534478361235e-06,
+      "loss": 0.2314,
+      "step": 12200
+    },
+    {
+      "epoch": 5.07698709945901,
+      "eval_main_loss": 0.2382633239030838,
+      "eval_main_runtime": 6.3506,
+      "eval_main_samples_per_second": 29.919,
+      "eval_main_steps_per_second": 3.779,
+      "step": 12200
+    },
+    {
+      "epoch": 5.07698709945901,
+      "eval_anatomy_loss": 2.834725856781006,
+      "eval_anatomy_runtime": 0.2681,
+      "eval_anatomy_samples_per_second": 7.461,
+      "eval_anatomy_steps_per_second": 3.73,
+      "step": 12200
+    },
+    {
+      "epoch": 5.07698709945901,
+      "eval_college_mathematics_loss": 2.052021026611328,
+      "eval_college_mathematics_runtime": 0.2663,
+      "eval_college_mathematics_samples_per_second": 7.511,
+      "eval_college_mathematics_steps_per_second": 3.756,
+      "step": 12200
+    },
+    {
+      "epoch": 5.07698709945901,
+      "eval_international_law_loss": 3.0643866062164307,
+      "eval_international_law_runtime": 0.2663,
+      "eval_international_law_samples_per_second": 7.51,
+      "eval_international_law_steps_per_second": 3.755,
+      "step": 12200
+    },
+    {
+      "epoch": 5.0853100291302535,
+      "grad_norm": 0.2890625,
+      "learning_rate": 3.3788086201306295e-06,
+      "loss": 0.2321,
+      "step": 12220
+    },
+    {
+      "epoch": 5.093632958801498,
+      "grad_norm": 0.263671875,
+      "learning_rate": 3.371154404575116e-06,
+      "loss": 0.2346,
+      "step": 12240
+    },
+    {
+      "epoch": 5.101955888472743,
+      "grad_norm": 0.33203125,
+      "learning_rate": 3.3634908829355384e-06,
+      "loss": 0.2298,
+      "step": 12260
+    },
+    {
+      "epoch": 5.110278818143986,
+      "grad_norm": 0.330078125,
+      "learning_rate": 3.3558181370772657e-06,
+      "loss": 0.2324,
+      "step": 12280
+    },
+    {
+      "epoch": 5.118601747815231,
+      "grad_norm": 0.236328125,
+      "learning_rate": 3.3481362489642055e-06,
+      "loss": 0.2314,
+      "step": 12300
+    },
+    {
+      "epoch": 5.118601747815231,
+      "eval_main_loss": 0.23837122321128845,
+      "eval_main_runtime": 6.354,
+      "eval_main_samples_per_second": 29.902,
+      "eval_main_steps_per_second": 3.777,
+      "step": 12300
+    },
+    {
+      "epoch": 5.118601747815231,
+      "eval_anatomy_loss": 2.833470582962036,
+      "eval_anatomy_runtime": 0.2671,
+      "eval_anatomy_samples_per_second": 7.488,
+      "eval_anatomy_steps_per_second": 3.744,
+      "step": 12300
+    },
+    {
+      "epoch": 5.118601747815231,
+      "eval_college_mathematics_loss": 2.0529050827026367,
+      "eval_college_mathematics_runtime": 0.2665,
+      "eval_college_mathematics_samples_per_second": 7.505,
+      "eval_college_mathematics_steps_per_second": 3.752,
+      "step": 12300
+    },
+    {
+      "epoch": 5.118601747815231,
+      "eval_international_law_loss": 3.0643563270568848,
+      "eval_international_law_runtime": 0.2685,
+      "eval_international_law_samples_per_second": 7.45,
+      "eval_international_law_steps_per_second": 3.725,
+      "step": 12300
+    },
+    {
+      "epoch": 5.126924677486476,
+      "grad_norm": 0.310546875,
+      "learning_rate": 3.340445300657924e-06,
+      "loss": 0.2306,
+      "step": 12320
+    },
+    {
+      "epoch": 5.135247607157719,
+      "grad_norm": 0.267578125,
+      "learning_rate": 3.3327453743167763e-06,
+      "loss": 0.2299,
+      "step": 12340
+    },
+    {
+      "epoch": 5.143570536828964,
+      "grad_norm": 0.27734375,
+      "learning_rate": 3.3250365521950212e-06,
+      "loss": 0.2334,
+      "step": 12360
+    },
+    {
+      "epoch": 5.1518934665002085,
+      "grad_norm": 0.24609375,
+      "learning_rate": 3.317318916641952e-06,
+      "loss": 0.2327,
+      "step": 12380
+    },
+    {
+      "epoch": 5.160216396171452,
+      "grad_norm": 0.267578125,
+      "learning_rate": 3.309592550101005e-06,
+      "loss": 0.2314,
+      "step": 12400
+    },
+    {
+      "epoch": 5.160216396171452,
+      "eval_main_loss": 0.2383902668952942,
+      "eval_main_runtime": 6.3193,
+      "eval_main_samples_per_second": 30.067,
+      "eval_main_steps_per_second": 3.798,
+      "step": 12400
+    },
+    {
+      "epoch": 5.160216396171452,
+      "eval_anatomy_loss": 2.8342981338500977,
+      "eval_anatomy_runtime": 0.2654,
+      "eval_anatomy_samples_per_second": 7.536,
+      "eval_anatomy_steps_per_second": 3.768,
+      "step": 12400
+    },
+    {
+      "epoch": 5.160216396171452,
+      "eval_college_mathematics_loss": 2.054227113723755,
+      "eval_college_mathematics_runtime": 0.2669,
+      "eval_college_mathematics_samples_per_second": 7.493,
+      "eval_college_mathematics_steps_per_second": 3.746,
+      "step": 12400
+    },
+    {
+      "epoch": 5.160216396171452,
+      "eval_international_law_loss": 3.067544937133789,
+      "eval_international_law_runtime": 0.2651,
+      "eval_international_law_samples_per_second": 7.543,
+      "eval_international_law_steps_per_second": 3.772,
+      "step": 12400
+    },
+    {
+      "epoch": 5.168539325842697,
+      "grad_norm": 0.2734375,
+      "learning_rate": 3.3018575351088894e-06,
+      "loss": 0.2336,
+      "step": 12420
+    },
+    {
+      "epoch": 5.1768622555139405,
+      "grad_norm": 0.310546875,
+      "learning_rate": 3.2941139542946996e-06,
+      "loss": 0.2368,
+      "step": 12440
+    },
+    {
+      "epoch": 5.185185185185185,
+      "grad_norm": 0.279296875,
+      "learning_rate": 3.2863618903790346e-06,
+      "loss": 0.2295,
+      "step": 12460
+    },
+    {
+      "epoch": 5.19350811485643,
+      "grad_norm": 0.267578125,
+      "learning_rate": 3.2786014261731138e-06,
+      "loss": 0.2322,
+      "step": 12480
+    },
+    {
+      "epoch": 5.201831044527673,
+      "grad_norm": 0.30859375,
+      "learning_rate": 3.270832644577891e-06,
+      "loss": 0.2365,
+      "step": 12500
+    },
+    {
+      "epoch": 5.201831044527673,
+      "eval_main_loss": 0.23834168910980225,
+      "eval_main_runtime": 6.3151,
+      "eval_main_samples_per_second": 30.087,
+      "eval_main_steps_per_second": 3.8,
+      "step": 12500
+    },
+    {
+      "epoch": 5.201831044527673,
+      "eval_anatomy_loss": 2.833904981613159,
+      "eval_anatomy_runtime": 0.2651,
+      "eval_anatomy_samples_per_second": 7.543,
+      "eval_anatomy_steps_per_second": 3.772,
+      "step": 12500
+    },
+    {
+      "epoch": 5.201831044527673,
+      "eval_college_mathematics_loss": 2.05311918258667,
+      "eval_college_mathematics_runtime": 0.2671,
+      "eval_college_mathematics_samples_per_second": 7.489,
+      "eval_college_mathematics_steps_per_second": 3.744,
+      "step": 12500
+    },
+    {
+      "epoch": 5.201831044527673,
+      "eval_international_law_loss": 3.065603494644165,
+      "eval_international_law_runtime": 0.2657,
+      "eval_international_law_samples_per_second": 7.527,
+      "eval_international_law_steps_per_second": 3.764,
+      "step": 12500
+    },
+    {
+      "epoch": 5.210153974198918,
+      "grad_norm": 0.26171875,
+      "learning_rate": 3.263055628583174e-06,
+      "loss": 0.2315,
+      "step": 12520
+    },
+    {
+      "epoch": 5.218476903870163,
+      "grad_norm": 0.33203125,
+      "learning_rate": 3.25527046126673e-06,
+      "loss": 0.2295,
+      "step": 12540
+    },
+    {
+      "epoch": 5.226799833541406,
+      "grad_norm": 0.271484375,
+      "learning_rate": 3.247477225793406e-06,
+      "loss": 0.2349,
+      "step": 12560
+    },
+    {
+      "epoch": 5.235122763212651,
+      "grad_norm": 0.26953125,
+      "learning_rate": 3.239676005414234e-06,
+      "loss": 0.231,
+      "step": 12580
+    },
+    {
+      "epoch": 5.2434456928838955,
+      "grad_norm": 0.251953125,
+      "learning_rate": 3.231866883465548e-06,
+      "loss": 0.2316,
+      "step": 12600
+    },
+    {
+      "epoch": 5.2434456928838955,
+      "eval_main_loss": 0.2384309619665146,
+      "eval_main_runtime": 6.3556,
+      "eval_main_samples_per_second": 29.895,
+      "eval_main_steps_per_second": 3.776,
+      "step": 12600
+    },
+    {
+      "epoch": 5.2434456928838955,
+      "eval_anatomy_loss": 2.8339719772338867,
+      "eval_anatomy_runtime": 0.266,
+      "eval_anatomy_samples_per_second": 7.518,
+      "eval_anatomy_steps_per_second": 3.759,
+      "step": 12600
+    },
+    {
+      "epoch": 5.2434456928838955,
+      "eval_college_mathematics_loss": 2.048271417617798,
+      "eval_college_mathematics_runtime": 0.2672,
+      "eval_college_mathematics_samples_per_second": 7.486,
+      "eval_college_mathematics_steps_per_second": 3.743,
+      "step": 12600
+    },
+    {
+      "epoch": 5.2434456928838955,
+      "eval_international_law_loss": 3.0657992362976074,
+      "eval_international_law_runtime": 0.2694,
+      "eval_international_law_samples_per_second": 7.425,
+      "eval_international_law_steps_per_second": 3.712,
+      "step": 12600
+    },
+    {
+      "epoch": 5.251768622555139,
+      "grad_norm": 0.28125,
+      "learning_rate": 3.2240499433680866e-06,
+      "loss": 0.2337,
+      "step": 12620
+    },
+    {
+      "epoch": 5.260091552226384,
+      "grad_norm": 0.275390625,
+      "learning_rate": 3.2162252686261077e-06,
+      "loss": 0.2284,
+      "step": 12640
+    },
+    {
+      "epoch": 5.2684144818976275,
+      "grad_norm": 0.302734375,
+      "learning_rate": 3.2083929428264938e-06,
+      "loss": 0.2318,
+      "step": 12660
+    },
+    {
+      "epoch": 5.276737411568872,
+      "grad_norm": 0.30078125,
+      "learning_rate": 3.2005530496378596e-06,
+      "loss": 0.2267,
+      "step": 12680
+    },
+    {
+      "epoch": 5.285060341240117,
+      "grad_norm": 0.271484375,
+      "learning_rate": 3.1927056728096582e-06,
+      "loss": 0.235,
+      "step": 12700
+    },
+    {
+      "epoch": 5.285060341240117,
+      "eval_main_loss": 0.23832783102989197,
+      "eval_main_runtime": 6.3472,
+      "eval_main_samples_per_second": 29.934,
+      "eval_main_steps_per_second": 3.781,
+      "step": 12700
+    },
+    {
+      "epoch": 5.285060341240117,
+      "eval_anatomy_loss": 2.834886312484741,
+      "eval_anatomy_runtime": 0.2666,
+      "eval_anatomy_samples_per_second": 7.503,
+      "eval_anatomy_steps_per_second": 3.751,
+      "step": 12700
+    },
+    {
+      "epoch": 5.285060341240117,
+      "eval_college_mathematics_loss": 2.0527701377868652,
+      "eval_college_mathematics_runtime": 0.2661,
+      "eval_college_mathematics_samples_per_second": 7.515,
+      "eval_college_mathematics_steps_per_second": 3.758,
+      "step": 12700
+    },
+    {
+      "epoch": 5.285060341240117,
+      "eval_international_law_loss": 3.0642306804656982,
+      "eval_international_law_runtime": 0.2678,
+      "eval_international_law_samples_per_second": 7.468,
+      "eval_international_law_steps_per_second": 3.734,
+      "step": 12700
+    },
+    {
+      "epoch": 5.29338327091136,
+      "grad_norm": 0.2734375,
+      "learning_rate": 3.184850896171288e-06,
+      "loss": 0.2332,
+      "step": 12720
+    },
+    {
+      "epoch": 5.301706200582605,
+      "grad_norm": 0.26171875,
+      "learning_rate": 3.1769888036311924e-06,
+      "loss": 0.2293,
+      "step": 12740
+    },
+    {
+      "epoch": 5.31002913025385,
+      "grad_norm": 0.26953125,
+      "learning_rate": 3.1691194791759693e-06,
+      "loss": 0.2336,
+      "step": 12760
+    },
+    {
+      "epoch": 5.318352059925093,
+      "grad_norm": 0.296875,
+      "learning_rate": 3.161243006869471e-06,
+      "loss": 0.2298,
+      "step": 12780
+    },
+    {
+      "epoch": 5.326674989596338,
+      "grad_norm": 0.3125,
+      "learning_rate": 3.1533594708519067e-06,
+      "loss": 0.2353,
+      "step": 12800
+    },
+    {
+      "epoch": 5.326674989596338,
+      "eval_main_loss": 0.23834091424942017,
+      "eval_main_runtime": 6.345,
+      "eval_main_samples_per_second": 29.945,
+      "eval_main_steps_per_second": 3.783,
+      "step": 12800
+    },
+    {
+      "epoch": 5.326674989596338,
+      "eval_anatomy_loss": 2.8317482471466064,
+      "eval_anatomy_runtime": 0.2676,
+      "eval_anatomy_samples_per_second": 7.474,
+      "eval_anatomy_steps_per_second": 3.737,
+      "step": 12800
+    },
+    {
+      "epoch": 5.326674989596338,
+      "eval_college_mathematics_loss": 2.0519726276397705,
+      "eval_college_mathematics_runtime": 0.2665,
+      "eval_college_mathematics_samples_per_second": 7.506,
+      "eval_college_mathematics_steps_per_second": 3.753,
+      "step": 12800
+    },
+    {
+      "epoch": 5.326674989596338,
+      "eval_international_law_loss": 3.0644750595092773,
+      "eval_international_law_runtime": 0.2671,
+      "eval_international_law_samples_per_second": 7.487,
+      "eval_international_law_steps_per_second": 3.744,
+      "step": 12800
+    },
+    {
+      "epoch": 5.3349979192675825,
+      "grad_norm": 0.279296875,
+      "learning_rate": 3.145468955338942e-06,
+      "loss": 0.2312,
+      "step": 12820
+    },
+    {
+      "epoch": 5.343320848938826,
+      "grad_norm": 0.279296875,
+      "learning_rate": 3.1375715446208014e-06,
+      "loss": 0.2297,
+      "step": 12840
+    },
+    {
+      "epoch": 5.351643778610071,
+      "grad_norm": 0.275390625,
+      "learning_rate": 3.129667323061369e-06,
+      "loss": 0.2341,
+      "step": 12860
+    },
+    {
+      "epoch": 5.359966708281315,
+      "grad_norm": 0.30859375,
+      "learning_rate": 3.1217563750972827e-06,
+      "loss": 0.232,
+      "step": 12880
+    },
+    {
+      "epoch": 5.368289637952559,
+      "grad_norm": 0.265625,
+      "learning_rate": 3.1138387852370385e-06,
+      "loss": 0.2255,
+      "step": 12900
+    },
+    {
+      "epoch": 5.368289637952559,
+      "eval_main_loss": 0.23833264410495758,
+      "eval_main_runtime": 6.3508,
+      "eval_main_samples_per_second": 29.917,
+      "eval_main_steps_per_second": 3.779,
+      "step": 12900
+    },
+    {
+      "epoch": 5.368289637952559,
+      "eval_anatomy_loss": 2.8347291946411133,
+      "eval_anatomy_runtime": 0.2672,
+      "eval_anatomy_samples_per_second": 7.484,
+      "eval_anatomy_steps_per_second": 3.742,
+      "step": 12900
+    },
+    {
+      "epoch": 5.368289637952559,
+      "eval_college_mathematics_loss": 2.050865411758423,
+      "eval_college_mathematics_runtime": 0.267,
+      "eval_college_mathematics_samples_per_second": 7.49,
+      "eval_college_mathematics_steps_per_second": 3.745,
+      "step": 12900
+    },
+    {
+      "epoch": 5.368289637952559,
+      "eval_international_law_loss": 3.068070411682129,
+      "eval_international_law_runtime": 0.2669,
+      "eval_international_law_samples_per_second": 7.494,
+      "eval_international_law_steps_per_second": 3.747,
+      "step": 12900
+    },
+    {
+      "epoch": 5.376612567623804,
+      "grad_norm": 0.27734375,
+      "learning_rate": 3.10591463806008e-06,
+      "loss": 0.2328,
+      "step": 12920
+    },
+    {
+      "epoch": 5.384935497295048,
+      "grad_norm": 0.29296875,
+      "learning_rate": 3.0979840182159033e-06,
+      "loss": 0.2346,
+      "step": 12940
+    },
+    {
+      "epoch": 5.393258426966292,
+      "grad_norm": 0.275390625,
+      "learning_rate": 3.0900470104231456e-06,
+      "loss": 0.2319,
+      "step": 12960
+    },
+    {
+      "epoch": 5.401581356637537,
+      "grad_norm": 0.2890625,
+      "learning_rate": 3.0821036994686837e-06,
+      "loss": 0.2343,
+      "step": 12980
+    },
+    {
+      "epoch": 5.40990428630878,
+      "grad_norm": 0.259765625,
+      "learning_rate": 3.074154170206731e-06,
+      "loss": 0.2299,
+      "step": 13000
+    },
+    {
+      "epoch": 5.40990428630878,
+      "eval_main_loss": 0.23825454711914062,
+      "eval_main_runtime": 6.3483,
+      "eval_main_samples_per_second": 29.929,
+      "eval_main_steps_per_second": 3.781,
+      "step": 13000
+    },
+    {
+      "epoch": 5.40990428630878,
+      "eval_anatomy_loss": 2.83435320854187,
+      "eval_anatomy_runtime": 0.2665,
+      "eval_anatomy_samples_per_second": 7.505,
+      "eval_anatomy_steps_per_second": 3.752,
+      "step": 13000
+    },
+    {
+      "epoch": 5.40990428630878,
+      "eval_college_mathematics_loss": 2.0521814823150635,
+      "eval_college_mathematics_runtime": 0.2672,
+      "eval_college_mathematics_samples_per_second": 7.485,
+      "eval_college_mathematics_steps_per_second": 3.743,
+      "step": 13000
+    },
+    {
+      "epoch": 5.40990428630878,
+      "eval_international_law_loss": 3.065732479095459,
+      "eval_international_law_runtime": 0.2669,
+      "eval_international_law_samples_per_second": 7.494,
+      "eval_international_law_steps_per_second": 3.747,
+      "step": 13000
+    },
+    {
+      "epoch": 5.418227215980025,
+      "grad_norm": 0.28125,
+      "learning_rate": 3.066198507557923e-06,
+      "loss": 0.2285,
+      "step": 13020
+    },
+    {
+      "epoch": 5.4265501456512695,
+      "grad_norm": 0.259765625,
+      "learning_rate": 3.0582367965084183e-06,
+      "loss": 0.2353,
+      "step": 13040
+    },
+    {
+      "epoch": 5.434873075322513,
+      "grad_norm": 0.328125,
+      "learning_rate": 3.0502691221089846e-06,
+      "loss": 0.2346,
+      "step": 13060
+    },
+    {
+      "epoch": 5.443196004993758,
+      "grad_norm": 0.27734375,
+      "learning_rate": 3.042295569474096e-06,
+      "loss": 0.2344,
+      "step": 13080
+    },
+    {
+      "epoch": 5.451518934665002,
+      "grad_norm": 0.287109375,
+      "learning_rate": 3.03431622378102e-06,
+      "loss": 0.2324,
+      "step": 13100
+    },
+    {
+      "epoch": 5.451518934665002,
+      "eval_main_loss": 0.23835337162017822,
+      "eval_main_runtime": 6.3367,
+      "eval_main_samples_per_second": 29.984,
+      "eval_main_steps_per_second": 3.787,
+      "step": 13100
+    },
+    {
+      "epoch": 5.451518934665002,
+      "eval_anatomy_loss": 2.835411787033081,
+      "eval_anatomy_runtime": 0.2674,
+      "eval_anatomy_samples_per_second": 7.479,
+      "eval_anatomy_steps_per_second": 3.74,
+      "step": 13100
+    },
+    {
+      "epoch": 5.451518934665002,
+      "eval_college_mathematics_loss": 2.0492405891418457,
+      "eval_college_mathematics_runtime": 0.2657,
+      "eval_college_mathematics_samples_per_second": 7.527,
+      "eval_college_mathematics_steps_per_second": 3.764,
+      "step": 13100
+    },
+    {
+      "epoch": 5.451518934665002,
+      "eval_international_law_loss": 3.0660183429718018,
+      "eval_international_law_runtime": 0.2667,
+      "eval_international_law_samples_per_second": 7.499,
+      "eval_international_law_steps_per_second": 3.75,
+      "step": 13100
+    },
+    {
+      "epoch": 5.459841864336246,
+      "grad_norm": 0.3125,
+      "learning_rate": 3.0263311702689045e-06,
+      "loss": 0.2356,
+      "step": 13120
+    },
+    {
+      "epoch": 5.468164794007491,
+      "grad_norm": 0.298828125,
+      "learning_rate": 3.018340494237878e-06,
+      "loss": 0.2302,
+      "step": 13140
+    },
+    {
+      "epoch": 5.476487723678735,
+      "grad_norm": 0.287109375,
+      "learning_rate": 3.010344281048125e-06,
+      "loss": 0.2322,
+      "step": 13160
+    },
+    {
+      "epoch": 5.484810653349979,
+      "grad_norm": 0.279296875,
+      "learning_rate": 3.0023426161189828e-06,
+      "loss": 0.2329,
+      "step": 13180
+    },
+    {
+      "epoch": 5.493133583021224,
+      "grad_norm": 0.31640625,
+      "learning_rate": 2.994335584928028e-06,
+      "loss": 0.233,
+      "step": 13200
+    },
+    {
+      "epoch": 5.493133583021224,
+      "eval_main_loss": 0.23826710879802704,
+      "eval_main_runtime": 6.3331,
+      "eval_main_samples_per_second": 30.001,
+      "eval_main_steps_per_second": 3.79,
+      "step": 13200
+    },
+    {
+      "epoch": 5.493133583021224,
+      "eval_anatomy_loss": 2.8339459896087646,
+      "eval_anatomy_runtime": 0.2666,
+      "eval_anatomy_samples_per_second": 7.501,
+      "eval_anatomy_steps_per_second": 3.751,
+      "step": 13200
+    },
+    {
+      "epoch": 5.493133583021224,
+      "eval_college_mathematics_loss": 2.051870107650757,
+      "eval_college_mathematics_runtime": 0.2669,
+      "eval_college_mathematics_samples_per_second": 7.494,
+      "eval_college_mathematics_steps_per_second": 3.747,
+      "step": 13200
+    },
+    {
+      "epoch": 5.493133583021224,
+      "eval_international_law_loss": 3.066417694091797,
+      "eval_international_law_runtime": 0.2661,
+      "eval_international_law_samples_per_second": 7.517,
+      "eval_international_law_steps_per_second": 3.758,
+      "step": 13200
+    },
+    {
+      "epoch": 5.501456512692467,
+      "grad_norm": 0.283203125,
+      "learning_rate": 2.9863232730101616e-06,
+      "loss": 0.231,
+      "step": 13220
+    },
+    {
+      "epoch": 5.509779442363712,
+      "grad_norm": 0.314453125,
+      "learning_rate": 2.9783057659566945e-06,
+      "loss": 0.2367,
+      "step": 13240
+    },
+    {
+      "epoch": 5.5181023720349565,
+      "grad_norm": 0.267578125,
+      "learning_rate": 2.9702831494144354e-06,
+      "loss": 0.2299,
+      "step": 13260
+    },
+    {
+      "epoch": 5.526425301706201,
+      "grad_norm": 0.3203125,
+      "learning_rate": 2.9622555090847756e-06,
+      "loss": 0.2303,
+      "step": 13280
+    },
+    {
+      "epoch": 5.534748231377445,
+      "grad_norm": 0.232421875,
+      "learning_rate": 2.954222930722771e-06,
+      "loss": 0.2281,
+      "step": 13300
+    },
+    {
+      "epoch": 5.534748231377445,
+      "eval_main_loss": 0.23840035498142242,
+      "eval_main_runtime": 6.3399,
+      "eval_main_samples_per_second": 29.969,
+      "eval_main_steps_per_second": 3.786,
+      "step": 13300
+    },
+    {
+      "epoch": 5.534748231377445,
+      "eval_anatomy_loss": 2.832589864730835,
+      "eval_anatomy_runtime": 0.2666,
+      "eval_anatomy_samples_per_second": 7.501,
+      "eval_anatomy_steps_per_second": 3.751,
+      "step": 13300
+    },
+    {
+      "epoch": 5.534748231377445,
+      "eval_college_mathematics_loss": 2.051084518432617,
+      "eval_college_mathematics_runtime": 0.2678,
+      "eval_college_mathematics_samples_per_second": 7.468,
+      "eval_college_mathematics_steps_per_second": 3.734,
+      "step": 13300
+    },
+    {
+      "epoch": 5.534748231377445,
+      "eval_international_law_loss": 3.0671823024749756,
+      "eval_international_law_runtime": 0.2664,
+      "eval_international_law_samples_per_second": 7.509,
+      "eval_international_law_steps_per_second": 3.754,
+      "step": 13300
+    },
+    {
+      "epoch": 5.543071161048689,
+      "grad_norm": 0.2734375,
+      "learning_rate": 2.9461855001362298e-06,
+      "loss": 0.2315,
+      "step": 13320
+    },
+    {
+      "epoch": 5.551394090719933,
+      "grad_norm": 0.2060546875,
+      "learning_rate": 2.9381433031847946e-06,
+      "loss": 0.2309,
+      "step": 13340
+    },
+    {
+      "epoch": 5.559717020391178,
+      "grad_norm": 0.294921875,
+      "learning_rate": 2.9300964257790215e-06,
+      "loss": 0.2326,
+      "step": 13360
+    },
+    {
+      "epoch": 5.568039950062422,
+      "grad_norm": 0.25,
+      "learning_rate": 2.9220449538794676e-06,
+      "loss": 0.2342,
+      "step": 13380
+    },
+    {
+      "epoch": 5.576362879733666,
+      "grad_norm": 0.3125,
+      "learning_rate": 2.9139889734957698e-06,
+      "loss": 0.23,
+      "step": 13400
+    },
+    {
+      "epoch": 5.576362879733666,
+      "eval_main_loss": 0.23834313452243805,
+      "eval_main_runtime": 6.3321,
+      "eval_main_samples_per_second": 30.006,
+      "eval_main_steps_per_second": 3.79,
+      "step": 13400
+    },
+    {
+      "epoch": 5.576362879733666,
+      "eval_anatomy_loss": 2.8331494331359863,
+      "eval_anatomy_runtime": 0.2663,
+      "eval_anatomy_samples_per_second": 7.512,
+      "eval_anatomy_steps_per_second": 3.756,
+      "step": 13400
+    },
+    {
+      "epoch": 5.576362879733666,
+      "eval_college_mathematics_loss": 2.0503456592559814,
+      "eval_college_mathematics_runtime": 0.2656,
+      "eval_college_mathematics_samples_per_second": 7.529,
+      "eval_college_mathematics_steps_per_second": 3.764,
+      "step": 13400
+    },
+    {
+      "epoch": 5.576362879733666,
+      "eval_international_law_loss": 3.065586805343628,
+      "eval_international_law_runtime": 0.2652,
+      "eval_international_law_samples_per_second": 7.542,
+      "eval_international_law_steps_per_second": 3.771,
+      "step": 13400
+    },
+    {
+      "epoch": 5.5846858094049106,
+      "grad_norm": 0.296875,
+      "learning_rate": 2.9059285706857287e-06,
+      "loss": 0.2328,
+      "step": 13420
+    },
+    {
+      "epoch": 5.593008739076155,
+      "grad_norm": 0.322265625,
+      "learning_rate": 2.897863831554385e-06,
+      "loss": 0.2294,
+      "step": 13440
+    },
+    {
+      "epoch": 5.601331668747399,
+      "grad_norm": 0.29296875,
+      "learning_rate": 2.889794842253102e-06,
+      "loss": 0.2362,
+      "step": 13460
+    },
+    {
+      "epoch": 5.6096545984186434,
+      "grad_norm": 0.337890625,
+      "learning_rate": 2.881721688978647e-06,
+      "loss": 0.2309,
+      "step": 13480
+    },
+    {
+      "epoch": 5.617977528089888,
+      "grad_norm": 0.279296875,
+      "learning_rate": 2.8736444579722665e-06,
+      "loss": 0.2347,
+      "step": 13500
+    },
+    {
+      "epoch": 5.617977528089888,
+      "eval_main_loss": 0.238357812166214,
+      "eval_main_runtime": 6.3227,
+      "eval_main_samples_per_second": 30.051,
+      "eval_main_steps_per_second": 3.796,
+      "step": 13500
+    },
+    {
+      "epoch": 5.617977528089888,
+      "eval_anatomy_loss": 2.834318161010742,
+      "eval_anatomy_runtime": 0.2663,
+      "eval_anatomy_samples_per_second": 7.511,
+      "eval_anatomy_steps_per_second": 3.756,
+      "step": 13500
+    },
+    {
+      "epoch": 5.617977528089888,
+      "eval_college_mathematics_loss": 2.0517024993896484,
+      "eval_college_mathematics_runtime": 0.2655,
+      "eval_college_mathematics_samples_per_second": 7.532,
+      "eval_college_mathematics_steps_per_second": 3.766,
+      "step": 13500
+    },
+    {
+      "epoch": 5.617977528089888,
+      "eval_international_law_loss": 3.0657360553741455,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.506,
+      "eval_international_law_steps_per_second": 3.753,
+      "step": 13500
+    },
+    {
+      "epoch": 5.626300457761132,
+      "grad_norm": 0.2890625,
+      "learning_rate": 2.865563235518772e-06,
+      "loss": 0.2299,
+      "step": 13520
+    },
+    {
+      "epoch": 5.634623387432376,
+      "grad_norm": 0.259765625,
+      "learning_rate": 2.8574781079456065e-06,
+      "loss": 0.2336,
+      "step": 13540
+    },
+    {
+      "epoch": 5.64294631710362,
+      "grad_norm": 0.251953125,
+      "learning_rate": 2.8493891616219354e-06,
+      "loss": 0.2347,
+      "step": 13560
+    },
+    {
+      "epoch": 5.651269246774865,
+      "grad_norm": 0.23046875,
+      "learning_rate": 2.841296482957715e-06,
+      "loss": 0.2323,
+      "step": 13580
+    },
+    {
+      "epoch": 5.659592176446109,
+      "grad_norm": 0.23046875,
+      "learning_rate": 2.8332001584027724e-06,
+      "loss": 0.233,
+      "step": 13600
+    },
+    {
+      "epoch": 5.659592176446109,
+      "eval_main_loss": 0.23819133639335632,
+      "eval_main_runtime": 6.333,
+      "eval_main_samples_per_second": 30.002,
+      "eval_main_steps_per_second": 3.79,
+      "step": 13600
+    },
+    {
+      "epoch": 5.659592176446109,
+      "eval_anatomy_loss": 2.8354241847991943,
+      "eval_anatomy_runtime": 0.2659,
+      "eval_anatomy_samples_per_second": 7.522,
+      "eval_anatomy_steps_per_second": 3.761,
+      "step": 13600
+    },
+    {
+      "epoch": 5.659592176446109,
+      "eval_college_mathematics_loss": 2.0492091178894043,
+      "eval_college_mathematics_runtime": 0.2664,
+      "eval_college_mathematics_samples_per_second": 7.508,
+      "eval_college_mathematics_steps_per_second": 3.754,
+      "step": 13600
+    },
+    {
+      "epoch": 5.659592176446109,
+      "eval_international_law_loss": 3.063929557800293,
+      "eval_international_law_runtime": 0.266,
+      "eval_international_law_samples_per_second": 7.52,
+      "eval_international_law_steps_per_second": 3.76,
+      "step": 13600
+    },
+    {
+      "epoch": 5.667915106117353,
+      "grad_norm": 0.2392578125,
+      "learning_rate": 2.825100274445882e-06,
+      "loss": 0.2308,
+      "step": 13620
+    },
+    {
+      "epoch": 5.6762380357885975,
+      "grad_norm": 0.2890625,
+      "learning_rate": 2.8169969176138435e-06,
+      "loss": 0.2305,
+      "step": 13640
+    },
+    {
+      "epoch": 5.684560965459842,
+      "grad_norm": 0.271484375,
+      "learning_rate": 2.808890174470551e-06,
+      "loss": 0.2316,
+      "step": 13660
+    },
+    {
+      "epoch": 5.692883895131086,
+      "grad_norm": 0.302734375,
+      "learning_rate": 2.8007801316160767e-06,
+      "loss": 0.2328,
+      "step": 13680
+    },
+    {
+      "epoch": 5.70120682480233,
+      "grad_norm": 0.2421875,
+      "learning_rate": 2.79266687568574e-06,
+      "loss": 0.2303,
+      "step": 13700
+    },
+    {
+      "epoch": 5.70120682480233,
+      "eval_main_loss": 0.23830586671829224,
+      "eval_main_runtime": 6.3319,
+      "eval_main_samples_per_second": 30.007,
+      "eval_main_steps_per_second": 3.79,
+      "step": 13700
+    },
+    {
+      "epoch": 5.70120682480233,
+      "eval_anatomy_loss": 2.833200216293335,
+      "eval_anatomy_runtime": 0.2674,
+      "eval_anatomy_samples_per_second": 7.48,
+      "eval_anatomy_steps_per_second": 3.74,
+      "step": 13700
+    },
+    {
+      "epoch": 5.70120682480233,
+      "eval_college_mathematics_loss": 2.051815986633301,
+      "eval_college_mathematics_runtime": 0.2651,
+      "eval_college_mathematics_samples_per_second": 7.543,
+      "eval_college_mathematics_steps_per_second": 3.772,
+      "step": 13700
+    },
+    {
+      "epoch": 5.70120682480233,
+      "eval_international_law_loss": 3.0680174827575684,
+      "eval_international_law_runtime": 0.2674,
+      "eval_international_law_samples_per_second": 7.478,
+      "eval_international_law_steps_per_second": 3.739,
+      "step": 13700
+    },
+    {
+      "epoch": 5.709529754473575,
+      "grad_norm": 0.25,
+      "learning_rate": 2.784550493349185e-06,
+      "loss": 0.2286,
+      "step": 13720
+    },
+    {
+      "epoch": 5.717852684144819,
+      "grad_norm": 0.271484375,
+      "learning_rate": 2.776431071309453e-06,
+      "loss": 0.227,
+      "step": 13740
+    },
+    {
+      "epoch": 5.726175613816063,
+      "grad_norm": 0.3046875,
+      "learning_rate": 2.7683086963020566e-06,
+      "loss": 0.2313,
+      "step": 13760
+    },
+    {
+      "epoch": 5.734498543487308,
+      "grad_norm": 0.25390625,
+      "learning_rate": 2.7601834550940538e-06,
+      "loss": 0.2275,
+      "step": 13780
+    },
+    {
+      "epoch": 5.742821473158552,
+      "grad_norm": 0.2490234375,
+      "learning_rate": 2.7520554344831194e-06,
+      "loss": 0.2313,
+      "step": 13800
+    },
+    {
+      "epoch": 5.742821473158552,
+      "eval_main_loss": 0.23832766711711884,
+      "eval_main_runtime": 6.3298,
+      "eval_main_samples_per_second": 30.017,
+      "eval_main_steps_per_second": 3.792,
+      "step": 13800
+    },
+    {
+      "epoch": 5.742821473158552,
+      "eval_anatomy_loss": 2.833317995071411,
+      "eval_anatomy_runtime": 0.2657,
+      "eval_anatomy_samples_per_second": 7.527,
+      "eval_anatomy_steps_per_second": 3.763,
+      "step": 13800
+    },
+    {
+      "epoch": 5.742821473158552,
+      "eval_college_mathematics_loss": 2.052196502685547,
+      "eval_college_mathematics_runtime": 0.2671,
+      "eval_college_mathematics_samples_per_second": 7.487,
+      "eval_college_mathematics_steps_per_second": 3.743,
+      "step": 13800
+    },
+    {
+      "epoch": 5.742821473158552,
+      "eval_international_law_loss": 3.066879987716675,
+      "eval_international_law_runtime": 0.2654,
+      "eval_international_law_samples_per_second": 7.536,
+      "eval_international_law_steps_per_second": 3.768,
+      "step": 13800
+    },
+    {
+      "epoch": 5.751144402829796,
+      "grad_norm": 0.2451171875,
+      "learning_rate": 2.743924721296622e-06,
+      "loss": 0.2323,
+      "step": 13820
+    },
+    {
+      "epoch": 5.759467332501041,
+      "grad_norm": 0.345703125,
+      "learning_rate": 2.735791402390691e-06,
+      "loss": 0.2325,
+      "step": 13840
+    },
+    {
+      "epoch": 5.7677902621722845,
+      "grad_norm": 0.275390625,
+      "learning_rate": 2.727655564649293e-06,
+      "loss": 0.2308,
+      "step": 13860
+    },
+    {
+      "epoch": 5.776113191843529,
+      "grad_norm": 0.30078125,
+      "learning_rate": 2.719517294983299e-06,
+      "loss": 0.2337,
+      "step": 13880
+    },
+    {
+      "epoch": 5.784436121514773,
+      "grad_norm": 0.2109375,
+      "learning_rate": 2.7113766803295637e-06,
+      "loss": 0.2319,
+      "step": 13900
+    },
+    {
+      "epoch": 5.784436121514773,
+      "eval_main_loss": 0.23832297325134277,
+      "eval_main_runtime": 6.3179,
+      "eval_main_samples_per_second": 30.073,
+      "eval_main_steps_per_second": 3.799,
+      "step": 13900
+    },
+    {
+      "epoch": 5.784436121514773,
+      "eval_anatomy_loss": 2.8318614959716797,
+      "eval_anatomy_runtime": 0.2657,
+      "eval_anatomy_samples_per_second": 7.526,
+      "eval_anatomy_steps_per_second": 3.763,
+      "step": 13900
+    },
+    {
+      "epoch": 5.784436121514773,
+      "eval_college_mathematics_loss": 2.0509073734283447,
+      "eval_college_mathematics_runtime": 0.2664,
+      "eval_college_mathematics_samples_per_second": 7.508,
+      "eval_college_mathematics_steps_per_second": 3.754,
+      "step": 13900
+    },
+    {
+      "epoch": 5.784436121514773,
+      "eval_international_law_loss": 3.0668725967407227,
+      "eval_international_law_runtime": 0.2662,
+      "eval_international_law_samples_per_second": 7.513,
+      "eval_international_law_steps_per_second": 3.756,
+      "step": 13900
+    },
+    {
+      "epoch": 5.792759051186017,
+      "grad_norm": 0.298828125,
+      "learning_rate": 2.7032338076499882e-06,
+      "loss": 0.2317,
+      "step": 13920
+    },
+    {
+      "epoch": 5.801081980857262,
+      "grad_norm": 0.2890625,
+      "learning_rate": 2.695088763930596e-06,
+      "loss": 0.2333,
+      "step": 13940
+    },
+    {
+      "epoch": 5.809404910528506,
+      "grad_norm": 0.2890625,
+      "learning_rate": 2.6869416361806026e-06,
+      "loss": 0.2296,
+      "step": 13960
+    },
+    {
+      "epoch": 5.81772784019975,
+      "grad_norm": 0.287109375,
+      "learning_rate": 2.6787925114314885e-06,
+      "loss": 0.2286,
+      "step": 13980
+    },
+    {
+      "epoch": 5.826050769870995,
+      "grad_norm": 0.283203125,
+      "learning_rate": 2.6706414767360615e-06,
+      "loss": 0.233,
+      "step": 14000
+    },
+    {
+      "epoch": 5.826050769870995,
+      "eval_main_loss": 0.23830385506153107,
+      "eval_main_runtime": 6.3332,
+      "eval_main_samples_per_second": 30.001,
+      "eval_main_steps_per_second": 3.79,
+      "step": 14000
+    },
+    {
+      "epoch": 5.826050769870995,
+      "eval_anatomy_loss": 2.8321759700775146,
+      "eval_anatomy_runtime": 0.2656,
+      "eval_anatomy_samples_per_second": 7.53,
+      "eval_anatomy_steps_per_second": 3.765,
+      "step": 14000
+    },
+    {
+      "epoch": 5.826050769870995,
+      "eval_college_mathematics_loss": 2.050740957260132,
+      "eval_college_mathematics_runtime": 0.2664,
+      "eval_college_mathematics_samples_per_second": 7.509,
+      "eval_college_mathematics_steps_per_second": 3.754,
+      "step": 14000
+    },
+    {
+      "epoch": 5.826050769870995,
+      "eval_international_law_loss": 3.068516492843628,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.506,
+      "eval_international_law_steps_per_second": 3.753,
+      "step": 14000
+    },
+    {
+      "epoch": 5.834373699542239,
+      "grad_norm": 0.2734375,
+      "learning_rate": 2.6624886191675387e-06,
+      "loss": 0.2309,
+      "step": 14020
+    },
+    {
+      "epoch": 5.842696629213483,
+      "grad_norm": 0.2392578125,
+      "learning_rate": 2.6543340258186063e-06,
+      "loss": 0.229,
+      "step": 14040
+    },
+    {
+      "epoch": 5.851019558884728,
+      "grad_norm": 0.2890625,
+      "learning_rate": 2.6461777838004933e-06,
+      "loss": 0.2307,
+      "step": 14060
+    },
+    {
+      "epoch": 5.8593424885559715,
+      "grad_norm": 0.2353515625,
+      "learning_rate": 2.6380199802420414e-06,
+      "loss": 0.2294,
+      "step": 14080
+    },
+    {
+      "epoch": 5.867665418227216,
+      "grad_norm": 0.298828125,
+      "learning_rate": 2.629860702288773e-06,
+      "loss": 0.234,
+      "step": 14100
+    },
+    {
+      "epoch": 5.867665418227216,
+      "eval_main_loss": 0.23834149539470673,
+      "eval_main_runtime": 6.3326,
+      "eval_main_samples_per_second": 30.003,
+      "eval_main_steps_per_second": 3.79,
+      "step": 14100
+    },
+    {
+      "epoch": 5.867665418227216,
+      "eval_anatomy_loss": 2.8320865631103516,
+      "eval_anatomy_runtime": 0.2669,
+      "eval_anatomy_samples_per_second": 7.493,
+      "eval_anatomy_steps_per_second": 3.747,
+      "step": 14100
+    },
+    {
+      "epoch": 5.867665418227216,
+      "eval_college_mathematics_loss": 2.049973249435425,
+      "eval_college_mathematics_runtime": 0.2663,
+      "eval_college_mathematics_samples_per_second": 7.509,
+      "eval_college_mathematics_steps_per_second": 3.755,
+      "step": 14100
+    },
+    {
+      "epoch": 5.867665418227216,
+      "eval_international_law_loss": 3.065117120742798,
+      "eval_international_law_runtime": 0.2661,
+      "eval_international_law_samples_per_second": 7.515,
+      "eval_international_law_steps_per_second": 3.758,
+      "step": 14100
+    },
+    {
+      "epoch": 5.87598834789846,
+      "grad_norm": 0.265625,
+      "learning_rate": 2.6217000371019597e-06,
+      "loss": 0.2286,
+      "step": 14120
+    },
+    {
+      "epoch": 5.884311277569704,
+      "grad_norm": 0.2060546875,
+      "learning_rate": 2.6135380718576947e-06,
+      "loss": 0.2321,
+      "step": 14140
+    },
+    {
+      "epoch": 5.892634207240949,
+      "grad_norm": 0.30859375,
+      "learning_rate": 2.6053748937459565e-06,
+      "loss": 0.2322,
+      "step": 14160
+    },
+    {
+      "epoch": 5.900957136912194,
+      "grad_norm": 0.25,
+      "learning_rate": 2.597210589969682e-06,
+      "loss": 0.2275,
+      "step": 14180
+    },
+    {
+      "epoch": 5.909280066583437,
+      "grad_norm": 0.318359375,
+      "learning_rate": 2.5890452477438318e-06,
+      "loss": 0.2323,
+      "step": 14200
+    },
+    {
+      "epoch": 5.909280066583437,
+      "eval_main_loss": 0.2382332980632782,
+      "eval_main_runtime": 6.3329,
+      "eval_main_samples_per_second": 30.002,
+      "eval_main_steps_per_second": 3.79,
+      "step": 14200
+    },
+    {
+      "epoch": 5.909280066583437,
+      "eval_anatomy_loss": 2.8326704502105713,
+      "eval_anatomy_runtime": 0.2666,
+      "eval_anatomy_samples_per_second": 7.503,
+      "eval_anatomy_steps_per_second": 3.751,
+      "step": 14200
+    },
+    {
+      "epoch": 5.909280066583437,
+      "eval_college_mathematics_loss": 2.0510854721069336,
+      "eval_college_mathematics_runtime": 0.2658,
+      "eval_college_mathematics_samples_per_second": 7.524,
+      "eval_college_mathematics_steps_per_second": 3.762,
+      "step": 14200
+    },
+    {
+      "epoch": 5.909280066583437,
+      "eval_international_law_loss": 3.0666379928588867,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.506,
+      "eval_international_law_steps_per_second": 3.753,
+      "step": 14200
+    },
+    {
+      "epoch": 5.917602996254682,
+      "grad_norm": 0.275390625,
+      "learning_rate": 2.5808789542944585e-06,
+      "loss": 0.2338,
+      "step": 14220
+    },
+    {
+      "epoch": 5.925925925925926,
+      "grad_norm": 0.29296875,
+      "learning_rate": 2.572711796857779e-06,
+      "loss": 0.2326,
+      "step": 14240
+    },
+    {
+      "epoch": 5.93424885559717,
+      "grad_norm": 0.25390625,
+      "learning_rate": 2.564543862679238e-06,
+      "loss": 0.2278,
+      "step": 14260
+    },
+    {
+      "epoch": 5.942571785268415,
+      "grad_norm": 0.279296875,
+      "learning_rate": 2.556375239012578e-06,
+      "loss": 0.2372,
+      "step": 14280
+    },
+    {
+      "epoch": 5.9508947149396585,
+      "grad_norm": 0.287109375,
+      "learning_rate": 2.5482060131189058e-06,
+      "loss": 0.232,
+      "step": 14300
+    },
+    {
+      "epoch": 5.9508947149396585,
+      "eval_main_loss": 0.2383614033460617,
+      "eval_main_runtime": 6.3311,
+      "eval_main_samples_per_second": 30.01,
+      "eval_main_steps_per_second": 3.791,
+      "step": 14300
+    },
+    {
+      "epoch": 5.9508947149396585,
+      "eval_anatomy_loss": 2.8360040187835693,
+      "eval_anatomy_runtime": 0.2658,
+      "eval_anatomy_samples_per_second": 7.524,
+      "eval_anatomy_steps_per_second": 3.762,
+      "step": 14300
+    },
+    {
+      "epoch": 5.9508947149396585,
+      "eval_college_mathematics_loss": 2.0509347915649414,
+      "eval_college_mathematics_runtime": 0.2653,
+      "eval_college_mathematics_samples_per_second": 7.538,
+      "eval_college_mathematics_steps_per_second": 3.769,
+      "step": 14300
+    },
+    {
+      "epoch": 5.9508947149396585,
+      "eval_international_law_loss": 3.064924478530884,
+      "eval_international_law_runtime": 0.2668,
+      "eval_international_law_samples_per_second": 7.497,
+      "eval_international_law_steps_per_second": 3.749,
+      "step": 14300
+    },
+    {
+      "epoch": 5.959217644610903,
+      "grad_norm": 0.27734375,
+      "learning_rate": 2.540036272265764e-06,
+      "loss": 0.2343,
+      "step": 14320
+    },
+    {
+      "epoch": 5.967540574282148,
+      "grad_norm": 0.283203125,
+      "learning_rate": 2.5318661037261955e-06,
+      "loss": 0.233,
+      "step": 14340
+    },
+    {
+      "epoch": 5.975863503953391,
+      "grad_norm": 0.2734375,
+      "learning_rate": 2.5236955947778096e-06,
+      "loss": 0.2315,
+      "step": 14360
+    },
+    {
+      "epoch": 5.984186433624636,
+      "grad_norm": 0.2890625,
+      "learning_rate": 2.515524832701854e-06,
+      "loss": 0.2357,
+      "step": 14380
+    },
+    {
+      "epoch": 5.992509363295881,
+      "grad_norm": 0.310546875,
+      "learning_rate": 2.507353904782281e-06,
+      "loss": 0.2323,
+      "step": 14400
+    },
+    {
+      "epoch": 5.992509363295881,
+      "eval_main_loss": 0.23834244906902313,
+      "eval_main_runtime": 6.3321,
+      "eval_main_samples_per_second": 30.006,
+      "eval_main_steps_per_second": 3.79,
+      "step": 14400
+    },
+    {
+      "epoch": 5.992509363295881,
+      "eval_anatomy_loss": 2.830299139022827,
+      "eval_anatomy_runtime": 0.266,
+      "eval_anatomy_samples_per_second": 7.52,
+      "eval_anatomy_steps_per_second": 3.76,
+      "step": 14400
+    },
+    {
+      "epoch": 5.992509363295881,
+      "eval_college_mathematics_loss": 2.0514543056488037,
+      "eval_college_mathematics_runtime": 0.2665,
+      "eval_college_mathematics_samples_per_second": 7.504,
+      "eval_college_mathematics_steps_per_second": 3.752,
+      "step": 14400
+    },
+    {
+      "epoch": 5.992509363295881,
+      "eval_international_law_loss": 3.0648834705352783,
+      "eval_international_law_runtime": 0.2653,
+      "eval_international_law_samples_per_second": 7.539,
+      "eval_international_law_steps_per_second": 3.769,
+      "step": 14400
+    },
+    {
+      "epoch": 6.000832292967124,
+      "grad_norm": 0.2890625,
+      "learning_rate": 2.4991828983048126e-06,
+      "loss": 0.2355,
+      "step": 14420
+    },
+    {
+      "epoch": 6.009155222638369,
+      "grad_norm": 0.296875,
+      "learning_rate": 2.4910119005560123e-06,
+      "loss": 0.2307,
+      "step": 14440
+    },
+    {
+      "epoch": 6.017478152309613,
+      "grad_norm": 0.306640625,
+      "learning_rate": 2.4828409988223487e-06,
+      "loss": 0.2313,
+      "step": 14460
+    },
+    {
+      "epoch": 6.025801081980857,
+      "grad_norm": 0.28125,
+      "learning_rate": 2.4746702803892637e-06,
+      "loss": 0.2315,
+      "step": 14480
+    },
+    {
+      "epoch": 6.034124011652102,
+      "grad_norm": 0.25390625,
+      "learning_rate": 2.4664998325402442e-06,
+      "loss": 0.2318,
+      "step": 14500
+    },
+    {
+      "epoch": 6.034124011652102,
+      "eval_main_loss": 0.23824182152748108,
+      "eval_main_runtime": 6.3272,
+      "eval_main_samples_per_second": 30.029,
+      "eval_main_steps_per_second": 3.793,
+      "step": 14500
+    },
+    {
+      "epoch": 6.034124011652102,
+      "eval_anatomy_loss": 2.8344292640686035,
+      "eval_anatomy_runtime": 0.2661,
+      "eval_anatomy_samples_per_second": 7.517,
+      "eval_anatomy_steps_per_second": 3.758,
+      "step": 14500
+    },
+    {
+      "epoch": 6.034124011652102,
+      "eval_college_mathematics_loss": 2.0524632930755615,
+      "eval_college_mathematics_runtime": 0.2661,
+      "eval_college_mathematics_samples_per_second": 7.517,
+      "eval_college_mathematics_steps_per_second": 3.759,
+      "step": 14500
+    },
+    {
+      "epoch": 6.034124011652102,
+      "eval_international_law_loss": 3.0627036094665527,
+      "eval_international_law_runtime": 0.2662,
+      "eval_international_law_samples_per_second": 7.514,
+      "eval_international_law_steps_per_second": 3.757,
+      "step": 14500
+    },
+    {
+      "epoch": 6.0424469413233455,
+      "grad_norm": 0.30078125,
+      "learning_rate": 2.4583297425558848e-06,
+      "loss": 0.2297,
+      "step": 14520
+    },
+    {
+      "epoch": 6.05076987099459,
+      "grad_norm": 0.29296875,
+      "learning_rate": 2.4501600977129564e-06,
+      "loss": 0.2293,
+      "step": 14540
+    },
+    {
+      "epoch": 6.059092800665835,
+      "grad_norm": 0.291015625,
+      "learning_rate": 2.441990985283476e-06,
+      "loss": 0.2329,
+      "step": 14560
+    },
+    {
+      "epoch": 6.067415730337078,
+      "grad_norm": 0.287109375,
+      "learning_rate": 2.433822492533774e-06,
+      "loss": 0.2308,
+      "step": 14580
+    },
+    {
+      "epoch": 6.075738660008323,
+      "grad_norm": 0.306640625,
+      "learning_rate": 2.4256547067235577e-06,
+      "loss": 0.2313,
+      "step": 14600
+    },
+    {
+      "epoch": 6.075738660008323,
+      "eval_main_loss": 0.2382703721523285,
+      "eval_main_runtime": 6.3262,
+      "eval_main_samples_per_second": 30.034,
+      "eval_main_steps_per_second": 3.794,
+      "step": 14600
+    },
+    {
+      "epoch": 6.075738660008323,
+      "eval_anatomy_loss": 2.832831859588623,
+      "eval_anatomy_runtime": 0.2652,
+      "eval_anatomy_samples_per_second": 7.543,
+      "eval_anatomy_steps_per_second": 3.771,
+      "step": 14600
+    },
+    {
+      "epoch": 6.075738660008323,
+      "eval_college_mathematics_loss": 2.0489039421081543,
+      "eval_college_mathematics_runtime": 0.2646,
+      "eval_college_mathematics_samples_per_second": 7.56,
+      "eval_college_mathematics_steps_per_second": 3.78,
+      "step": 14600
+    },
+    {
+      "epoch": 6.075738660008323,
+      "eval_international_law_loss": 3.067589044570923,
+      "eval_international_law_runtime": 0.2662,
+      "eval_international_law_samples_per_second": 7.514,
+      "eval_international_law_steps_per_second": 3.757,
+      "step": 14600
+    },
+    {
+      "epoch": 6.084061589679568,
+      "grad_norm": 0.25,
+      "learning_rate": 2.4174877151049852e-06,
+      "loss": 0.2373,
+      "step": 14620
+    },
+    {
+      "epoch": 6.092384519350811,
+      "grad_norm": 0.275390625,
+      "learning_rate": 2.4093216049217315e-06,
+      "loss": 0.2279,
+      "step": 14640
+    },
+    {
+      "epoch": 6.100707449022056,
+      "grad_norm": 0.291015625,
+      "learning_rate": 2.4011564634080527e-06,
+      "loss": 0.2341,
+      "step": 14660
+    },
+    {
+      "epoch": 6.1090303786933005,
+      "grad_norm": 0.345703125,
+      "learning_rate": 2.3929923777878596e-06,
+      "loss": 0.2316,
+      "step": 14680
+    },
+    {
+      "epoch": 6.117353308364544,
+      "grad_norm": 0.25390625,
+      "learning_rate": 2.3848294352737837e-06,
+      "loss": 0.2363,
+      "step": 14700
+    },
+    {
+      "epoch": 6.117353308364544,
+      "eval_main_loss": 0.23824332654476166,
+      "eval_main_runtime": 6.3273,
+      "eval_main_samples_per_second": 30.028,
+      "eval_main_steps_per_second": 3.793,
+      "step": 14700
+    },
+    {
+      "epoch": 6.117353308364544,
+      "eval_anatomy_loss": 2.832658290863037,
+      "eval_anatomy_runtime": 0.2668,
+      "eval_anatomy_samples_per_second": 7.497,
+      "eval_anatomy_steps_per_second": 3.748,
+      "step": 14700
+    },
+    {
+      "epoch": 6.117353308364544,
+      "eval_college_mathematics_loss": 2.05214524269104,
+      "eval_college_mathematics_runtime": 0.2655,
+      "eval_college_mathematics_samples_per_second": 7.532,
+      "eval_college_mathematics_steps_per_second": 3.766,
+      "step": 14700
+    },
+    {
+      "epoch": 6.117353308364544,
+      "eval_international_law_loss": 3.068110942840576,
+      "eval_international_law_runtime": 0.2657,
+      "eval_international_law_samples_per_second": 7.528,
+      "eval_international_law_steps_per_second": 3.764,
+      "step": 14700
+    },
+    {
+      "epoch": 6.125676238035789,
+      "grad_norm": 0.31640625,
+      "learning_rate": 2.3766677230662413e-06,
+      "loss": 0.2342,
+      "step": 14720
+    },
+    {
+      "epoch": 6.1339991677070325,
+      "grad_norm": 0.267578125,
+      "learning_rate": 2.368507328352511e-06,
+      "loss": 0.2281,
+      "step": 14740
+    },
+    {
+      "epoch": 6.142322097378277,
+      "grad_norm": 0.263671875,
+      "learning_rate": 2.3603483383057944e-06,
+      "loss": 0.2332,
+      "step": 14760
+    },
+    {
+      "epoch": 6.150645027049522,
+      "grad_norm": 0.27734375,
+      "learning_rate": 2.352190840084288e-06,
+      "loss": 0.2308,
+      "step": 14780
+    },
+    {
+      "epoch": 6.158967956720765,
+      "grad_norm": 0.30859375,
+      "learning_rate": 2.3440349208302553e-06,
+      "loss": 0.2364,
+      "step": 14800
+    },
+    {
+      "epoch": 6.158967956720765,
+      "eval_main_loss": 0.2383221834897995,
+      "eval_main_runtime": 6.332,
+      "eval_main_samples_per_second": 30.006,
+      "eval_main_steps_per_second": 3.79,
+      "step": 14800
+    },
+    {
+      "epoch": 6.158967956720765,
+      "eval_anatomy_loss": 2.832653045654297,
+      "eval_anatomy_runtime": 0.2664,
+      "eval_anatomy_samples_per_second": 7.506,
+      "eval_anatomy_steps_per_second": 3.753,
+      "step": 14800
+    },
+    {
+      "epoch": 6.158967956720765,
+      "eval_college_mathematics_loss": 2.0538275241851807,
+      "eval_college_mathematics_runtime": 0.2663,
+      "eval_college_mathematics_samples_per_second": 7.509,
+      "eval_college_mathematics_steps_per_second": 3.755,
+      "step": 14800
+    },
+    {
+      "epoch": 6.158967956720765,
+      "eval_international_law_loss": 3.0637497901916504,
+      "eval_international_law_runtime": 0.2654,
+      "eval_international_law_samples_per_second": 7.536,
+      "eval_international_law_steps_per_second": 3.768,
+      "step": 14800
+    },
+    {
+      "epoch": 6.16729088639201,
+      "grad_norm": 0.26171875,
+      "learning_rate": 2.3358806676690855e-06,
+      "loss": 0.2316,
+      "step": 14820
+    },
+    {
+      "epoch": 6.1756138160632545,
+      "grad_norm": 0.265625,
+      "learning_rate": 2.327728167708377e-06,
+      "loss": 0.2321,
+      "step": 14840
+    },
+    {
+      "epoch": 6.183936745734498,
+      "grad_norm": 0.2109375,
+      "learning_rate": 2.3195775080369954e-06,
+      "loss": 0.2358,
+      "step": 14860
+    },
+    {
+      "epoch": 6.192259675405743,
+      "grad_norm": 0.296875,
+      "learning_rate": 2.3114287757241487e-06,
+      "loss": 0.2296,
+      "step": 14880
+    },
+    {
+      "epoch": 6.200582605076987,
+      "grad_norm": 0.29296875,
+      "learning_rate": 2.3032820578184567e-06,
+      "loss": 0.2284,
+      "step": 14900
+    },
+    {
+      "epoch": 6.200582605076987,
+      "eval_main_loss": 0.23826871812343597,
+      "eval_main_runtime": 6.3212,
+      "eval_main_samples_per_second": 30.058,
+      "eval_main_steps_per_second": 3.797,
+      "step": 14900
+    },
+    {
+      "epoch": 6.200582605076987,
+      "eval_anatomy_loss": 2.8343729972839355,
+      "eval_anatomy_runtime": 0.2658,
+      "eval_anatomy_samples_per_second": 7.526,
+      "eval_anatomy_steps_per_second": 3.763,
+      "step": 14900
+    },
+    {
+      "epoch": 6.200582605076987,
+      "eval_college_mathematics_loss": 2.051208257675171,
+      "eval_college_mathematics_runtime": 0.266,
+      "eval_college_mathematics_samples_per_second": 7.52,
+      "eval_college_mathematics_steps_per_second": 3.76,
+      "step": 14900
+    },
+    {
+      "epoch": 6.200582605076987,
+      "eval_international_law_loss": 3.064528465270996,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.505,
+      "eval_international_law_steps_per_second": 3.753,
+      "step": 14900
+    },
+    {
+      "epoch": 6.208905534748231,
+      "grad_norm": 0.2490234375,
+      "learning_rate": 2.29513744134702e-06,
+      "loss": 0.2272,
+      "step": 14920
+    },
+    {
+      "epoch": 6.217228464419476,
+      "grad_norm": 0.3046875,
+      "learning_rate": 2.286995013314488e-06,
+      "loss": 0.2331,
+      "step": 14940
+    },
+    {
+      "epoch": 6.22555139409072,
+      "grad_norm": 0.318359375,
+      "learning_rate": 2.2788548607021366e-06,
+      "loss": 0.2334,
+      "step": 14960
+    },
+    {
+      "epoch": 6.233874323761964,
+      "grad_norm": 0.2734375,
+      "learning_rate": 2.270717070466933e-06,
+      "loss": 0.2327,
+      "step": 14980
+    },
+    {
+      "epoch": 6.242197253433209,
+      "grad_norm": 0.306640625,
+      "learning_rate": 2.262581729540605e-06,
+      "loss": 0.2337,
+      "step": 15000
+    },
+    {
+      "epoch": 6.242197253433209,
+      "eval_main_loss": 0.238285630941391,
+      "eval_main_runtime": 6.3306,
+      "eval_main_samples_per_second": 30.013,
+      "eval_main_steps_per_second": 3.791,
+      "step": 15000
+    },
+    {
+      "epoch": 6.242197253433209,
+      "eval_anatomy_loss": 2.83296537399292,
+      "eval_anatomy_runtime": 0.2659,
+      "eval_anatomy_samples_per_second": 7.521,
+      "eval_anatomy_steps_per_second": 3.761,
+      "step": 15000
+    },
+    {
+      "epoch": 6.242197253433209,
+      "eval_college_mathematics_loss": 2.051861047744751,
+      "eval_college_mathematics_runtime": 0.2655,
+      "eval_college_mathematics_samples_per_second": 7.532,
+      "eval_college_mathematics_steps_per_second": 3.766,
+      "step": 15000
+    },
+    {
+      "epoch": 6.242197253433209,
+      "eval_international_law_loss": 3.066511631011963,
+      "eval_international_law_runtime": 0.2683,
+      "eval_international_law_samples_per_second": 7.455,
+      "eval_international_law_steps_per_second": 3.728,
+      "step": 15000
+    },
+    {
+      "epoch": 6.250520183104452,
+      "grad_norm": 0.291015625,
+      "learning_rate": 2.2544489248287218e-06,
+      "loss": 0.2295,
+      "step": 15020
+    },
+    {
+      "epoch": 6.258843112775697,
+      "grad_norm": 0.27734375,
+      "learning_rate": 2.246318743209753e-06,
+      "loss": 0.2312,
+      "step": 15040
+    },
+    {
+      "epoch": 6.2671660424469415,
+      "grad_norm": 0.2421875,
+      "learning_rate": 2.2381912715341528e-06,
+      "loss": 0.2327,
+      "step": 15060
+    },
+    {
+      "epoch": 6.275488972118185,
+      "grad_norm": 0.2890625,
+      "learning_rate": 2.2300665966234243e-06,
+      "loss": 0.2317,
+      "step": 15080
+    },
+    {
+      "epoch": 6.28381190178943,
+      "grad_norm": 0.30078125,
+      "learning_rate": 2.221944805269192e-06,
+      "loss": 0.2274,
+      "step": 15100
+    },
+    {
+      "epoch": 6.28381190178943,
+      "eval_main_loss": 0.23841296136379242,
+      "eval_main_runtime": 6.3599,
+      "eval_main_samples_per_second": 29.875,
+      "eval_main_steps_per_second": 3.774,
+      "step": 15100
+    },
+    {
+      "epoch": 6.28381190178943,
+      "eval_anatomy_loss": 2.8354194164276123,
+      "eval_anatomy_runtime": 0.2665,
+      "eval_anatomy_samples_per_second": 7.505,
+      "eval_anatomy_steps_per_second": 3.753,
+      "step": 15100
+    },
+    {
+      "epoch": 6.28381190178943,
+      "eval_college_mathematics_loss": 2.0525858402252197,
+      "eval_college_mathematics_runtime": 0.2681,
+      "eval_college_mathematics_samples_per_second": 7.461,
+      "eval_college_mathematics_steps_per_second": 3.73,
+      "step": 15100
+    },
+    {
+      "epoch": 6.28381190178943,
+      "eval_international_law_loss": 3.0661587715148926,
+      "eval_international_law_runtime": 0.2672,
+      "eval_international_law_samples_per_second": 7.484,
+      "eval_international_law_steps_per_second": 3.742,
+      "step": 15100
+    },
+    {
+      "epoch": 6.292134831460674,
+      "grad_norm": 0.267578125,
+      "learning_rate": 2.2138259842322794e-06,
+      "loss": 0.2311,
+      "step": 15120
+    },
+    {
+      "epoch": 6.300457761131918,
+      "grad_norm": 0.3203125,
+      "learning_rate": 2.2057102202417806e-06,
+      "loss": 0.2308,
+      "step": 15140
+    },
+    {
+      "epoch": 6.308780690803163,
+      "grad_norm": 0.28125,
+      "learning_rate": 2.1975975999941298e-06,
+      "loss": 0.232,
+      "step": 15160
+    },
+    {
+      "epoch": 6.317103620474407,
+      "grad_norm": 0.259765625,
+      "learning_rate": 2.1894882101521807e-06,
+      "loss": 0.2317,
+      "step": 15180
+    },
+    {
+      "epoch": 6.325426550145651,
+      "grad_norm": 0.275390625,
+      "learning_rate": 2.181382137344278e-06,
+      "loss": 0.2302,
+      "step": 15200
+    },
+    {
+      "epoch": 6.325426550145651,
+      "eval_main_loss": 0.23836155235767365,
+      "eval_main_runtime": 6.3527,
+      "eval_main_samples_per_second": 29.909,
+      "eval_main_steps_per_second": 3.778,
+      "step": 15200
+    },
+    {
+      "epoch": 6.325426550145651,
+      "eval_anatomy_loss": 2.8317906856536865,
+      "eval_anatomy_runtime": 0.2665,
+      "eval_anatomy_samples_per_second": 7.506,
+      "eval_anatomy_steps_per_second": 3.753,
+      "step": 15200
+    },
+    {
+      "epoch": 6.325426550145651,
+      "eval_college_mathematics_loss": 2.054029703140259,
+      "eval_college_mathematics_runtime": 0.2679,
+      "eval_college_mathematics_samples_per_second": 7.467,
+      "eval_college_mathematics_steps_per_second": 3.733,
+      "step": 15200
+    },
+    {
+      "epoch": 6.325426550145651,
+      "eval_international_law_loss": 3.0642471313476562,
+      "eval_international_law_runtime": 0.2663,
+      "eval_international_law_samples_per_second": 7.511,
+      "eval_international_law_steps_per_second": 3.755,
+      "step": 15200
+    },
+    {
+      "epoch": 6.333749479816896,
+      "grad_norm": 0.26171875,
+      "learning_rate": 2.173279468163331e-06,
+      "loss": 0.2322,
+      "step": 15220
+    },
+    {
+      "epoch": 6.34207240948814,
+      "grad_norm": 0.291015625,
+      "learning_rate": 2.1651802891658897e-06,
+      "loss": 0.2321,
+      "step": 15240
+    },
+    {
+      "epoch": 6.350395339159384,
+      "grad_norm": 0.345703125,
+      "learning_rate": 2.1570846868712227e-06,
+      "loss": 0.2378,
+      "step": 15260
+    },
+    {
+      "epoch": 6.3587182688306285,
+      "grad_norm": 0.25,
+      "learning_rate": 2.1489927477603884e-06,
+      "loss": 0.2301,
+      "step": 15280
+    },
+    {
+      "epoch": 6.367041198501873,
+      "grad_norm": 0.26953125,
+      "learning_rate": 2.1409045582753144e-06,
+      "loss": 0.234,
+      "step": 15300
+    },
+    {
+      "epoch": 6.367041198501873,
+      "eval_main_loss": 0.23826418817043304,
+      "eval_main_runtime": 6.3363,
+      "eval_main_samples_per_second": 29.986,
+      "eval_main_steps_per_second": 3.788,
+      "step": 15300
+    },
+    {
+      "epoch": 6.367041198501873,
+      "eval_anatomy_loss": 2.8337223529815674,
+      "eval_anatomy_runtime": 0.2673,
+      "eval_anatomy_samples_per_second": 7.483,
+      "eval_anatomy_steps_per_second": 3.741,
+      "step": 15300
+    },
+    {
+      "epoch": 6.367041198501873,
+      "eval_college_mathematics_loss": 2.0507359504699707,
+      "eval_college_mathematics_runtime": 0.2666,
+      "eval_college_mathematics_samples_per_second": 7.503,
+      "eval_college_mathematics_steps_per_second": 3.751,
+      "step": 15300
+    },
+    {
+      "epoch": 6.367041198501873,
+      "eval_international_law_loss": 3.065744161605835,
+      "eval_international_law_runtime": 0.2653,
+      "eval_international_law_samples_per_second": 7.538,
+      "eval_international_law_steps_per_second": 3.769,
+      "step": 15300
+    },
+    {
+      "epoch": 6.375364128173117,
+      "grad_norm": 0.318359375,
+      "learning_rate": 2.132820204817872e-06,
+      "loss": 0.2279,
+      "step": 15320
+    },
+    {
+      "epoch": 6.383687057844361,
+      "grad_norm": 0.251953125,
+      "learning_rate": 2.124739773748955e-06,
+      "loss": 0.2297,
+      "step": 15340
+    },
+    {
+      "epoch": 6.392009987515605,
+      "grad_norm": 0.275390625,
+      "learning_rate": 2.1166633513875563e-06,
+      "loss": 0.2296,
+      "step": 15360
+    },
+    {
+      "epoch": 6.40033291718685,
+      "grad_norm": 0.27734375,
+      "learning_rate": 2.1085910240098456e-06,
+      "loss": 0.2306,
+      "step": 15380
+    },
+    {
+      "epoch": 6.408655846858094,
+      "grad_norm": 0.2890625,
+      "learning_rate": 2.1005228778482484e-06,
+      "loss": 0.2319,
+      "step": 15400
+    },
+    {
+      "epoch": 6.408655846858094,
+      "eval_main_loss": 0.238305002450943,
+      "eval_main_runtime": 6.3303,
+      "eval_main_samples_per_second": 30.014,
+      "eval_main_steps_per_second": 3.791,
+      "step": 15400
+    },
+    {
+      "epoch": 6.408655846858094,
+      "eval_anatomy_loss": 2.8330578804016113,
+      "eval_anatomy_runtime": 0.2654,
+      "eval_anatomy_samples_per_second": 7.535,
+      "eval_anatomy_steps_per_second": 3.768,
+      "step": 15400
+    },
+    {
+      "epoch": 6.408655846858094,
+      "eval_college_mathematics_loss": 2.0521047115325928,
+      "eval_college_mathematics_runtime": 0.2667,
+      "eval_college_mathematics_samples_per_second": 7.498,
+      "eval_college_mathematics_steps_per_second": 3.749,
+      "step": 15400
+    },
+    {
+      "epoch": 6.408655846858094,
+      "eval_international_law_loss": 3.0654828548431396,
+      "eval_international_law_runtime": 0.4213,
+      "eval_international_law_samples_per_second": 4.748,
+      "eval_international_law_steps_per_second": 2.374,
+      "step": 15400
+    },
+    {
+      "epoch": 6.416978776529338,
+      "grad_norm": 0.265625,
+      "learning_rate": 2.0924589990905253e-06,
+      "loss": 0.2338,
+      "step": 15420
+    },
+    {
+      "epoch": 6.425301706200583,
+      "grad_norm": 0.2734375,
+      "learning_rate": 2.084399473878848e-06,
+      "loss": 0.2353,
+      "step": 15440
+    },
+    {
+      "epoch": 6.433624635871827,
+      "grad_norm": 0.28515625,
+      "learning_rate": 2.0763443883088833e-06,
+      "loss": 0.2326,
+      "step": 15460
+    },
+    {
+      "epoch": 6.441947565543071,
+      "grad_norm": 0.2255859375,
+      "learning_rate": 2.068293828428872e-06,
+      "loss": 0.2361,
+      "step": 15480
+    },
+    {
+      "epoch": 6.4502704952143155,
+      "grad_norm": 0.33203125,
+      "learning_rate": 2.0602478802387085e-06,
+      "loss": 0.2288,
+      "step": 15500
+    },
+    {
+      "epoch": 6.4502704952143155,
+      "eval_main_loss": 0.23823831975460052,
+      "eval_main_runtime": 6.3318,
+      "eval_main_samples_per_second": 30.007,
+      "eval_main_steps_per_second": 3.79,
+      "step": 15500
+    },
+    {
+      "epoch": 6.4502704952143155,
+      "eval_anatomy_loss": 2.834176540374756,
+      "eval_anatomy_runtime": 0.2663,
+      "eval_anatomy_samples_per_second": 7.51,
+      "eval_anatomy_steps_per_second": 3.755,
+      "step": 15500
+    },
+    {
+      "epoch": 6.4502704952143155,
+      "eval_college_mathematics_loss": 2.052133560180664,
+      "eval_college_mathematics_runtime": 0.2676,
+      "eval_college_mathematics_samples_per_second": 7.473,
+      "eval_college_mathematics_steps_per_second": 3.736,
+      "step": 15500
+    },
+    {
+      "epoch": 6.4502704952143155,
+      "eval_international_law_loss": 3.066138982772827,
+      "eval_international_law_runtime": 0.4176,
+      "eval_international_law_samples_per_second": 4.789,
+      "eval_international_law_steps_per_second": 2.394,
+      "step": 15500
+    },
+    {
+      "epoch": 6.45859342488556,
+      "grad_norm": 0.29296875,
+      "learning_rate": 2.0522066296890226e-06,
+      "loss": 0.2292,
+      "step": 15520
+    },
+    {
+      "epoch": 6.466916354556804,
+      "grad_norm": 0.314453125,
+      "learning_rate": 2.0441701626802647e-06,
+      "loss": 0.2313,
+      "step": 15540
+    },
+    {
+      "epoch": 6.475239284228048,
+      "grad_norm": 0.29296875,
+      "learning_rate": 2.036138565061779e-06,
+      "loss": 0.2322,
+      "step": 15560
+    },
+    {
+      "epoch": 6.483562213899292,
+      "grad_norm": 0.287109375,
+      "learning_rate": 2.0281119226308976e-06,
+      "loss": 0.2331,
+      "step": 15580
+    },
+    {
+      "epoch": 6.491885143570537,
+      "grad_norm": 0.26953125,
+      "learning_rate": 2.02009032113202e-06,
+      "loss": 0.232,
+      "step": 15600
+    },
+    {
+      "epoch": 6.491885143570537,
+      "eval_main_loss": 0.2383803129196167,
+      "eval_main_runtime": 6.3247,
+      "eval_main_samples_per_second": 30.041,
+      "eval_main_steps_per_second": 3.795,
+      "step": 15600
+    },
+    {
+      "epoch": 6.491885143570537,
+      "eval_anatomy_loss": 2.8346643447875977,
+      "eval_anatomy_runtime": 0.2658,
+      "eval_anatomy_samples_per_second": 7.524,
+      "eval_anatomy_steps_per_second": 3.762,
+      "step": 15600
+    },
+    {
+      "epoch": 6.491885143570537,
+      "eval_college_mathematics_loss": 2.0503625869750977,
+      "eval_college_mathematics_runtime": 0.2647,
+      "eval_college_mathematics_samples_per_second": 7.554,
+      "eval_college_mathematics_steps_per_second": 3.777,
+      "step": 15600
+    },
+    {
+      "epoch": 6.491885143570537,
+      "eval_international_law_loss": 3.066617012023926,
+      "eval_international_law_runtime": 0.2652,
+      "eval_international_law_samples_per_second": 7.542,
+      "eval_international_law_steps_per_second": 3.771,
+      "step": 15600
+    },
+    {
+      "epoch": 6.500208073241781,
+      "grad_norm": 0.271484375,
+      "learning_rate": 2.012073846255691e-06,
+      "loss": 0.2364,
+      "step": 15620
+    },
+    {
+      "epoch": 6.508531002913025,
+      "grad_norm": 0.291015625,
+      "learning_rate": 2.0040625836376937e-06,
+      "loss": 0.2308,
+      "step": 15640
+    },
+    {
+      "epoch": 6.51685393258427,
+      "grad_norm": 0.28515625,
+      "learning_rate": 1.9960566188581306e-06,
+      "loss": 0.2287,
+      "step": 15660
+    },
+    {
+      "epoch": 6.525176862255514,
+      "grad_norm": 0.3203125,
+      "learning_rate": 1.9880560374405107e-06,
+      "loss": 0.2299,
+      "step": 15680
+    },
+    {
+      "epoch": 6.533499791926758,
+      "grad_norm": 0.29296875,
+      "learning_rate": 1.980060924850836e-06,
+      "loss": 0.2325,
+      "step": 15700
+    },
+    {
+      "epoch": 6.533499791926758,
+      "eval_main_loss": 0.2383701652288437,
+      "eval_main_runtime": 6.3198,
+      "eval_main_samples_per_second": 30.064,
+      "eval_main_steps_per_second": 3.798,
+      "step": 15700
+    },
+    {
+      "epoch": 6.533499791926758,
+      "eval_anatomy_loss": 2.833786725997925,
+      "eval_anatomy_runtime": 0.2657,
+      "eval_anatomy_samples_per_second": 7.528,
+      "eval_anatomy_steps_per_second": 3.764,
+      "step": 15700
+    },
+    {
+      "epoch": 6.533499791926758,
+      "eval_college_mathematics_loss": 2.050478219985962,
+      "eval_college_mathematics_runtime": 0.2678,
+      "eval_college_mathematics_samples_per_second": 7.467,
+      "eval_college_mathematics_steps_per_second": 3.734,
+      "step": 15700
+    },
+    {
+      "epoch": 6.533499791926758,
+      "eval_international_law_loss": 3.06551194190979,
+      "eval_international_law_runtime": 0.2654,
+      "eval_international_law_samples_per_second": 7.535,
+      "eval_international_law_steps_per_second": 3.767,
+      "step": 15700
+    },
+    {
+      "epoch": 6.5418227215980025,
+      "grad_norm": 0.283203125,
+      "learning_rate": 1.972071366496685e-06,
+      "loss": 0.2301,
+      "step": 15720
+    },
+    {
+      "epoch": 6.550145651269247,
+      "grad_norm": 0.287109375,
+      "learning_rate": 1.964087447726306e-06,
+      "loss": 0.2336,
+      "step": 15740
+    },
+    {
+      "epoch": 6.558468580940491,
+      "grad_norm": 0.275390625,
+      "learning_rate": 1.956109253827702e-06,
+      "loss": 0.2355,
+      "step": 15760
+    },
+    {
+      "epoch": 6.566791510611735,
+      "grad_norm": 0.302734375,
+      "learning_rate": 1.9481368700277197e-06,
+      "loss": 0.2309,
+      "step": 15780
+    },
+    {
+      "epoch": 6.57511444028298,
+      "grad_norm": 0.267578125,
+      "learning_rate": 1.9401703814911394e-06,
+      "loss": 0.2334,
+      "step": 15800
+    },
+    {
+      "epoch": 6.57511444028298,
+      "eval_main_loss": 0.2383488267660141,
+      "eval_main_runtime": 6.3287,
+      "eval_main_samples_per_second": 30.022,
+      "eval_main_steps_per_second": 3.792,
+      "step": 15800
+    },
+    {
+      "epoch": 6.57511444028298,
+      "eval_anatomy_loss": 2.8328428268432617,
+      "eval_anatomy_runtime": 0.2649,
+      "eval_anatomy_samples_per_second": 7.551,
+      "eval_anatomy_steps_per_second": 3.776,
+      "step": 15800
+    },
+    {
+      "epoch": 6.57511444028298,
+      "eval_college_mathematics_loss": 2.052703619003296,
+      "eval_college_mathematics_runtime": 0.2652,
+      "eval_college_mathematics_samples_per_second": 7.54,
+      "eval_college_mathematics_steps_per_second": 3.77,
+      "step": 15800
+    },
+    {
+      "epoch": 6.57511444028298,
+      "eval_international_law_loss": 3.065504550933838,
+      "eval_international_law_runtime": 0.2679,
+      "eval_international_law_samples_per_second": 7.466,
+      "eval_international_law_steps_per_second": 3.733,
+      "step": 15800
+    },
+    {
+      "epoch": 6.583437369954224,
+      "grad_norm": 0.2578125,
+      "learning_rate": 1.9322098733197677e-06,
+      "loss": 0.2323,
+      "step": 15820
+    },
+    {
+      "epoch": 6.591760299625468,
+      "grad_norm": 0.283203125,
+      "learning_rate": 1.9242554305515208e-06,
+      "loss": 0.2327,
+      "step": 15840
+    },
+    {
+      "epoch": 6.600083229296713,
+      "grad_norm": 0.322265625,
+      "learning_rate": 1.9163071381595273e-06,
+      "loss": 0.2335,
+      "step": 15860
+    },
+    {
+      "epoch": 6.608406158967957,
+      "grad_norm": 0.2373046875,
+      "learning_rate": 1.908365081051212e-06,
+      "loss": 0.2286,
+      "step": 15880
+    },
+    {
+      "epoch": 6.616729088639201,
+      "grad_norm": 0.35546875,
+      "learning_rate": 1.9004293440673908e-06,
+      "loss": 0.2317,
+      "step": 15900
+    },
+    {
+      "epoch": 6.616729088639201,
+      "eval_main_loss": 0.23834270238876343,
+      "eval_main_runtime": 6.3247,
+      "eval_main_samples_per_second": 30.041,
+      "eval_main_steps_per_second": 3.795,
+      "step": 15900
+    },
+    {
+      "epoch": 6.616729088639201,
+      "eval_anatomy_loss": 2.832962989807129,
+      "eval_anatomy_runtime": 0.2669,
+      "eval_anatomy_samples_per_second": 7.492,
+      "eval_anatomy_steps_per_second": 3.746,
+      "step": 15900
+    },
+    {
+      "epoch": 6.616729088639201,
+      "eval_college_mathematics_loss": 2.0485119819641113,
+      "eval_college_mathematics_runtime": 0.2668,
+      "eval_college_mathematics_samples_per_second": 7.495,
+      "eval_college_mathematics_steps_per_second": 3.747,
+      "step": 15900
+    },
+    {
+      "epoch": 6.616729088639201,
+      "eval_international_law_loss": 3.066624164581299,
+      "eval_international_law_runtime": 0.2657,
+      "eval_international_law_samples_per_second": 7.527,
+      "eval_international_law_steps_per_second": 3.763,
+      "step": 15900
+    },
+    {
+      "epoch": 6.625052018310445,
+      "grad_norm": 0.31640625,
+      "learning_rate": 1.8925000119813669e-06,
+      "loss": 0.2292,
+      "step": 15920
+    },
+    {
+      "epoch": 6.6333749479816895,
+      "grad_norm": 0.251953125,
+      "learning_rate": 1.88457716949802e-06,
+      "loss": 0.231,
+      "step": 15940
+    },
+    {
+      "epoch": 6.641697877652934,
+      "grad_norm": 0.3828125,
+      "learning_rate": 1.8766609012529085e-06,
+      "loss": 0.2292,
+      "step": 15960
+    },
+    {
+      "epoch": 6.650020807324178,
+      "grad_norm": 0.2236328125,
+      "learning_rate": 1.8687512918113604e-06,
+      "loss": 0.2347,
+      "step": 15980
+    },
+    {
+      "epoch": 6.658343736995422,
+      "grad_norm": 0.28515625,
+      "learning_rate": 1.8608484256675702e-06,
+      "loss": 0.2292,
+      "step": 16000
+    },
+    {
+      "epoch": 6.658343736995422,
+      "eval_main_loss": 0.23837390542030334,
+      "eval_main_runtime": 6.3517,
+      "eval_main_samples_per_second": 29.913,
+      "eval_main_steps_per_second": 3.779,
+      "step": 16000
+    },
+    {
+      "epoch": 6.658343736995422,
+      "eval_anatomy_loss": 2.8328123092651367,
+      "eval_anatomy_runtime": 0.2671,
+      "eval_anatomy_samples_per_second": 7.488,
+      "eval_anatomy_steps_per_second": 3.744,
+      "step": 16000
+    },
+    {
+      "epoch": 6.658343736995422,
+      "eval_college_mathematics_loss": 2.0518500804901123,
+      "eval_college_mathematics_runtime": 0.2671,
+      "eval_college_mathematics_samples_per_second": 7.489,
+      "eval_college_mathematics_steps_per_second": 3.745,
+      "step": 16000
+    },
+    {
+      "epoch": 6.658343736995422,
+      "eval_international_law_loss": 3.0677542686462402,
+      "eval_international_law_runtime": 0.2658,
+      "eval_international_law_samples_per_second": 7.523,
+      "eval_international_law_steps_per_second": 3.762,
+      "step": 16000
+    },
+    {
+      "epoch": 6.666666666666667,
+      "grad_norm": 0.244140625,
+      "learning_rate": 1.852952387243698e-06,
+      "loss": 0.2263,
+      "step": 16020
+    },
+    {
+      "epoch": 6.674989596337911,
+      "grad_norm": 0.21484375,
+      "learning_rate": 1.845063260888969e-06,
+      "loss": 0.2312,
+      "step": 16040
+    },
+    {
+      "epoch": 6.683312526009155,
+      "grad_norm": 0.298828125,
+      "learning_rate": 1.8371811308787655e-06,
+      "loss": 0.2325,
+      "step": 16060
+    },
+    {
+      "epoch": 6.6916354556804,
+      "grad_norm": 0.24609375,
+      "learning_rate": 1.829306081413736e-06,
+      "loss": 0.2292,
+      "step": 16080
+    },
+    {
+      "epoch": 6.699958385351644,
+      "grad_norm": 0.328125,
+      "learning_rate": 1.8214381966188898e-06,
+      "loss": 0.2314,
+      "step": 16100
+    },
+    {
+      "epoch": 6.699958385351644,
+      "eval_main_loss": 0.2383338063955307,
+      "eval_main_runtime": 6.3513,
+      "eval_main_samples_per_second": 29.915,
+      "eval_main_steps_per_second": 3.779,
+      "step": 16100
+    },
+    {
+      "epoch": 6.699958385351644,
+      "eval_anatomy_loss": 2.8324191570281982,
+      "eval_anatomy_runtime": 0.2674,
+      "eval_anatomy_samples_per_second": 7.479,
+      "eval_anatomy_steps_per_second": 3.74,
+      "step": 16100
+    },
+    {
+      "epoch": 6.699958385351644,
+      "eval_college_mathematics_loss": 2.0509135723114014,
+      "eval_college_mathematics_runtime": 0.2664,
+      "eval_college_mathematics_samples_per_second": 7.507,
+      "eval_college_mathematics_steps_per_second": 3.754,
+      "step": 16100
+    },
+    {
+      "epoch": 6.699958385351644,
+      "eval_international_law_loss": 3.066457748413086,
+      "eval_international_law_runtime": 0.2667,
+      "eval_international_law_samples_per_second": 7.499,
+      "eval_international_law_steps_per_second": 3.749,
+      "step": 16100
+    },
+    {
+      "epoch": 6.708281315022888,
+      "grad_norm": 0.2890625,
+      "learning_rate": 1.813577560542699e-06,
+      "loss": 0.2288,
+      "step": 16120
+    },
+    {
+      "epoch": 6.716604244694132,
+      "grad_norm": 0.28515625,
+      "learning_rate": 1.8057242571562034e-06,
+      "loss": 0.2331,
+      "step": 16140
+    },
+    {
+      "epoch": 6.7249271743653765,
+      "grad_norm": 0.302734375,
+      "learning_rate": 1.7978783703521102e-06,
+      "loss": 0.2371,
+      "step": 16160
+    },
+    {
+      "epoch": 6.733250104036621,
+      "grad_norm": 0.283203125,
+      "learning_rate": 1.7900399839438976e-06,
+      "loss": 0.236,
+      "step": 16180
+    },
+    {
+      "epoch": 6.741573033707866,
+      "grad_norm": 0.298828125,
+      "learning_rate": 1.782209181664924e-06,
+      "loss": 0.2352,
+      "step": 16200
+    },
+    {
+      "epoch": 6.741573033707866,
+      "eval_main_loss": 0.23834922909736633,
+      "eval_main_runtime": 6.3393,
+      "eval_main_samples_per_second": 29.972,
+      "eval_main_steps_per_second": 3.786,
+      "step": 16200
+    },
+    {
+      "epoch": 6.741573033707866,
+      "eval_anatomy_loss": 2.832639217376709,
+      "eval_anatomy_runtime": 0.2655,
+      "eval_anatomy_samples_per_second": 7.533,
+      "eval_anatomy_steps_per_second": 3.766,
+      "step": 16200
+    },
+    {
+      "epoch": 6.741573033707866,
+      "eval_college_mathematics_loss": 2.050755023956299,
+      "eval_college_mathematics_runtime": 0.2658,
+      "eval_college_mathematics_samples_per_second": 7.525,
+      "eval_college_mathematics_steps_per_second": 3.762,
+      "step": 16200
+    },
+    {
+      "epoch": 6.741573033707866,
+      "eval_international_law_loss": 3.0636000633239746,
+      "eval_international_law_runtime": 0.2666,
+      "eval_international_law_samples_per_second": 7.502,
+      "eval_international_law_steps_per_second": 3.751,
+      "step": 16200
+    },
+    {
+      "epoch": 6.749895963379109,
+      "grad_norm": 0.291015625,
+      "learning_rate": 1.774386047167529e-06,
+      "loss": 0.2301,
+      "step": 16220
+    },
+    {
+      "epoch": 6.758218893050354,
+      "grad_norm": 0.275390625,
+      "learning_rate": 1.7665706640221415e-06,
+      "loss": 0.236,
+      "step": 16240
+    },
+    {
+      "epoch": 6.766541822721598,
+      "grad_norm": 0.287109375,
+      "learning_rate": 1.7587631157163876e-06,
+      "loss": 0.2328,
+      "step": 16260
+    },
+    {
+      "epoch": 6.774864752392842,
+      "grad_norm": 0.296875,
+      "learning_rate": 1.7509634856541951e-06,
+      "loss": 0.2337,
+      "step": 16280
+    },
+    {
+      "epoch": 6.783187682064087,
+      "grad_norm": 0.283203125,
+      "learning_rate": 1.7431718571549092e-06,
+      "loss": 0.231,
+      "step": 16300
+    },
+    {
+      "epoch": 6.783187682064087,
+      "eval_main_loss": 0.23830153048038483,
+      "eval_main_runtime": 6.3187,
+      "eval_main_samples_per_second": 30.069,
+      "eval_main_steps_per_second": 3.798,
+      "step": 16300
+    },
+    {
+      "epoch": 6.783187682064087,
+      "eval_anatomy_loss": 2.8300514221191406,
+      "eval_anatomy_runtime": 0.267,
+      "eval_anatomy_samples_per_second": 7.49,
+      "eval_anatomy_steps_per_second": 3.745,
+      "step": 16300
+    },
+    {
+      "epoch": 6.783187682064087,
+      "eval_college_mathematics_loss": 2.052163600921631,
+      "eval_college_mathematics_runtime": 0.2643,
+      "eval_college_mathematics_samples_per_second": 7.568,
+      "eval_college_mathematics_steps_per_second": 3.784,
+      "step": 16300
+    },
+    {
+      "epoch": 6.783187682064087,
+      "eval_international_law_loss": 3.064903736114502,
+      "eval_international_law_runtime": 0.2653,
+      "eval_international_law_samples_per_second": 7.539,
+      "eval_international_law_steps_per_second": 3.77,
+      "step": 16300
+    },
+    {
+      "epoch": 6.7915106117353305,
+      "grad_norm": 0.26171875,
+      "learning_rate": 1.7353883134523975e-06,
+      "loss": 0.231,
+      "step": 16320
+    },
+    {
+      "epoch": 6.799833541406575,
+      "grad_norm": 0.259765625,
+      "learning_rate": 1.7276129376941594e-06,
+      "loss": 0.2344,
+      "step": 16340
+    },
+    {
+      "epoch": 6.80815647107782,
+      "grad_norm": 0.265625,
+      "learning_rate": 1.7198458129404433e-06,
+      "loss": 0.2312,
+      "step": 16360
+    },
+    {
+      "epoch": 6.8164794007490634,
+      "grad_norm": 0.2470703125,
+      "learning_rate": 1.7120870221633556e-06,
+      "loss": 0.2315,
+      "step": 16380
+    },
+    {
+      "epoch": 6.824802330420308,
+      "grad_norm": 0.251953125,
+      "learning_rate": 1.704336648245975e-06,
+      "loss": 0.2272,
+      "step": 16400
+    },
+    {
+      "epoch": 6.824802330420308,
+      "eval_main_loss": 0.23840200901031494,
+      "eval_main_runtime": 6.3137,
+      "eval_main_samples_per_second": 30.093,
+      "eval_main_steps_per_second": 3.801,
+      "step": 16400
+    },
+    {
+      "epoch": 6.824802330420308,
+      "eval_anatomy_loss": 2.8353002071380615,
+      "eval_anatomy_runtime": 0.2656,
+      "eval_anatomy_samples_per_second": 7.53,
+      "eval_anatomy_steps_per_second": 3.765,
+      "step": 16400
+    },
+    {
+      "epoch": 6.824802330420308,
+      "eval_college_mathematics_loss": 2.0492591857910156,
+      "eval_college_mathematics_runtime": 0.2643,
+      "eval_college_mathematics_samples_per_second": 7.568,
+      "eval_college_mathematics_steps_per_second": 3.784,
+      "step": 16400
+    },
+    {
+      "epoch": 6.824802330420308,
+      "eval_international_law_loss": 3.065768241882324,
+      "eval_international_law_runtime": 0.2656,
+      "eval_international_law_samples_per_second": 7.531,
+      "eval_international_law_steps_per_second": 3.765,
+      "step": 16400
+    },
+    {
+      "epoch": 6.833125260091553,
+      "grad_norm": 0.271484375,
+      "learning_rate": 1.6965947739814672e-06,
+      "loss": 0.2333,
+      "step": 16420
+    },
+    {
+      "epoch": 6.841448189762796,
+      "grad_norm": 0.298828125,
+      "learning_rate": 1.6888614820722024e-06,
+      "loss": 0.2353,
+      "step": 16440
+    },
+    {
+      "epoch": 6.849771119434041,
+      "grad_norm": 0.279296875,
+      "learning_rate": 1.681136855128866e-06,
+      "loss": 0.2247,
+      "step": 16460
+    },
+    {
+      "epoch": 6.858094049105285,
+      "grad_norm": 0.279296875,
+      "learning_rate": 1.6734209756695843e-06,
+      "loss": 0.2363,
+      "step": 16480
+    },
+    {
+      "epoch": 6.866416978776529,
+      "grad_norm": 0.255859375,
+      "learning_rate": 1.6657139261190364e-06,
+      "loss": 0.2328,
+      "step": 16500
+    },
+    {
+      "epoch": 6.866416978776529,
+      "eval_main_loss": 0.23840853571891785,
+      "eval_main_runtime": 6.3197,
+      "eval_main_samples_per_second": 30.065,
+      "eval_main_steps_per_second": 3.798,
+      "step": 16500
+    },
+    {
+      "epoch": 6.866416978776529,
+      "eval_anatomy_loss": 2.834096908569336,
+      "eval_anatomy_runtime": 0.2667,
+      "eval_anatomy_samples_per_second": 7.5,
+      "eval_anatomy_steps_per_second": 3.75,
+      "step": 16500
+    },
+    {
+      "epoch": 6.866416978776529,
+      "eval_college_mathematics_loss": 2.0504603385925293,
+      "eval_college_mathematics_runtime": 0.265,
+      "eval_college_mathematics_samples_per_second": 7.546,
+      "eval_college_mathematics_steps_per_second": 3.773,
+      "step": 16500
+    },
+    {
+      "epoch": 6.866416978776529,
+      "eval_international_law_loss": 3.065997838973999,
+      "eval_international_law_runtime": 0.2658,
+      "eval_international_law_samples_per_second": 7.525,
+      "eval_international_law_steps_per_second": 3.762,
+      "step": 16500
+    },
+    {
+      "epoch": 6.874739908447774,
+      "grad_norm": 0.322265625,
+      "learning_rate": 1.6580157888075766e-06,
+      "loss": 0.2319,
+      "step": 16520
+    },
+    {
+      "epoch": 6.8830628381190175,
+      "grad_norm": 0.2578125,
+      "learning_rate": 1.6503266459703566e-06,
+      "loss": 0.2363,
+      "step": 16540
+    },
+    {
+      "epoch": 6.891385767790262,
+      "grad_norm": 0.287109375,
+      "learning_rate": 1.6426465797464402e-06,
+      "loss": 0.2318,
+      "step": 16560
+    },
+    {
+      "epoch": 6.899708697461507,
+      "grad_norm": 0.23828125,
+      "learning_rate": 1.6349756721779348e-06,
+      "loss": 0.2304,
+      "step": 16580
+    },
+    {
+      "epoch": 6.90803162713275,
+      "grad_norm": 0.28125,
+      "learning_rate": 1.6273140052091097e-06,
+      "loss": 0.2345,
+      "step": 16600
+    },
+    {
+      "epoch": 6.90803162713275,
+      "eval_main_loss": 0.2384311705827713,
+      "eval_main_runtime": 6.3138,
+      "eval_main_samples_per_second": 30.093,
+      "eval_main_steps_per_second": 3.801,
+      "step": 16600
+    },
+    {
+      "epoch": 6.90803162713275,
+      "eval_anatomy_loss": 2.832714319229126,
+      "eval_anatomy_runtime": 0.2643,
+      "eval_anatomy_samples_per_second": 7.566,
+      "eval_anatomy_steps_per_second": 3.783,
+      "step": 16600
+    },
+    {
+      "epoch": 6.90803162713275,
+      "eval_college_mathematics_loss": 2.0529298782348633,
+      "eval_college_mathematics_runtime": 0.2651,
+      "eval_college_mathematics_samples_per_second": 7.543,
+      "eval_college_mathematics_steps_per_second": 3.772,
+      "step": 16600
+    },
+    {
+      "epoch": 6.90803162713275,
+      "eval_international_law_loss": 3.064915180206299,
+      "eval_international_law_runtime": 0.2668,
+      "eval_international_law_samples_per_second": 7.497,
+      "eval_international_law_steps_per_second": 3.748,
+      "step": 16600
+    },
+    {
+      "epoch": 6.916354556803995,
+      "grad_norm": 0.306640625,
+      "learning_rate": 1.6196616606855194e-06,
+      "loss": 0.231,
+      "step": 16620
+    },
+    {
+      "epoch": 6.92467748647524,
+      "grad_norm": 0.265625,
+      "learning_rate": 1.6120187203531351e-06,
+      "loss": 0.2332,
+      "step": 16640
+    },
+    {
+      "epoch": 6.933000416146483,
+      "grad_norm": 0.2451171875,
+      "learning_rate": 1.6043852658574666e-06,
+      "loss": 0.2305,
+      "step": 16660
+    },
+    {
+      "epoch": 6.941323345817728,
+      "grad_norm": 0.318359375,
+      "learning_rate": 1.596761378742689e-06,
+      "loss": 0.2346,
+      "step": 16680
+    },
+    {
+      "epoch": 6.949646275488972,
+      "grad_norm": 0.291015625,
+      "learning_rate": 1.589147140450778e-06,
+      "loss": 0.2263,
+      "step": 16700
+    },
+    {
+      "epoch": 6.949646275488972,
+      "eval_main_loss": 0.23828622698783875,
+      "eval_main_runtime": 6.3112,
+      "eval_main_samples_per_second": 30.105,
+      "eval_main_steps_per_second": 3.803,
+      "step": 16700
+    },
+    {
+      "epoch": 6.949646275488972,
+      "eval_anatomy_loss": 2.833873748779297,
+      "eval_anatomy_runtime": 0.2668,
+      "eval_anatomy_samples_per_second": 7.497,
+      "eval_anatomy_steps_per_second": 3.748,
+      "step": 16700
+    },
+    {
+      "epoch": 6.949646275488972,
+      "eval_college_mathematics_loss": 2.0528018474578857,
+      "eval_college_mathematics_runtime": 0.2655,
+      "eval_college_mathematics_samples_per_second": 7.534,
+      "eval_college_mathematics_steps_per_second": 3.767,
+      "step": 16700
+    },
+    {
+      "epoch": 6.949646275488972,
+      "eval_international_law_loss": 3.0665361881256104,
+      "eval_international_law_runtime": 0.2655,
+      "eval_international_law_samples_per_second": 7.533,
+      "eval_international_law_steps_per_second": 3.766,
+      "step": 16700
+    },
+    {
+      "epoch": 6.957969205160216,
+      "grad_norm": 0.271484375,
+      "learning_rate": 1.5815426323206345e-06,
+      "loss": 0.2324,
+      "step": 16720
+    },
+    {
+      "epoch": 6.966292134831461,
+      "grad_norm": 0.2421875,
+      "learning_rate": 1.5739479355872162e-06,
+      "loss": 0.2307,
+      "step": 16740
+    },
+    {
+      "epoch": 6.974615064502705,
+      "grad_norm": 0.2421875,
+      "learning_rate": 1.5663631313806726e-06,
+      "loss": 0.235,
+      "step": 16760
+    },
+    {
+      "epoch": 6.982937994173949,
+      "grad_norm": 0.212890625,
+      "learning_rate": 1.5587883007254741e-06,
+      "loss": 0.2266,
+      "step": 16780
+    },
+    {
+      "epoch": 6.991260923845194,
+      "grad_norm": 0.265625,
+      "learning_rate": 1.5512235245395514e-06,
+      "loss": 0.2316,
+      "step": 16800
+    },
+    {
+      "epoch": 6.991260923845194,
+      "eval_main_loss": 0.23832102119922638,
+      "eval_main_runtime": 6.3531,
+      "eval_main_samples_per_second": 29.907,
+      "eval_main_steps_per_second": 3.778,
+      "step": 16800
+    },
+    {
+      "epoch": 6.991260923845194,
+      "eval_anatomy_loss": 2.8318288326263428,
+      "eval_anatomy_runtime": 0.2673,
+      "eval_anatomy_samples_per_second": 7.483,
+      "eval_anatomy_steps_per_second": 3.741,
+      "step": 16800
+    },
+    {
+      "epoch": 6.991260923845194,
+      "eval_college_mathematics_loss": 2.0499653816223145,
+      "eval_college_mathematics_runtime": 0.2665,
+      "eval_college_mathematics_samples_per_second": 7.505,
+      "eval_college_mathematics_steps_per_second": 3.752,
+      "step": 16800
+    },
+    {
+      "epoch": 6.991260923845194,
+      "eval_international_law_loss": 3.06691312789917,
+      "eval_international_law_runtime": 0.2681,
+      "eval_international_law_samples_per_second": 7.46,
+      "eval_international_law_steps_per_second": 3.73,
+      "step": 16800
+    },
+    {
+      "epoch": 6.999583853516437,
+      "grad_norm": 0.251953125,
+      "learning_rate": 1.5436688836334274e-06,
+      "loss": 0.2314,
+      "step": 16820
+    },
+    {
+      "epoch": 7.007906783187682,
+      "grad_norm": 0.294921875,
+      "learning_rate": 1.5361244587093551e-06,
+      "loss": 0.2344,
+      "step": 16840
+    },
+    {
+      "epoch": 7.016229712858927,
+      "grad_norm": 0.2421875,
+      "learning_rate": 1.528590330360456e-06,
+      "loss": 0.2314,
+      "step": 16860
+    },
+    {
+      "epoch": 7.02455264253017,
+      "grad_norm": 0.2890625,
+      "learning_rate": 1.5210665790698592e-06,
+      "loss": 0.2334,
+      "step": 16880
+    },
+    {
+      "epoch": 7.032875572201415,
+      "grad_norm": 0.279296875,
+      "learning_rate": 1.513553285209838e-06,
+      "loss": 0.2308,
+      "step": 16900
+    },
+    {
+      "epoch": 7.032875572201415,
+      "eval_main_loss": 0.2382553517818451,
+      "eval_main_runtime": 6.3599,
+      "eval_main_samples_per_second": 29.875,
+      "eval_main_steps_per_second": 3.774,
+      "step": 16900
+    },
+    {
+      "epoch": 7.032875572201415,
+      "eval_anatomy_loss": 2.832185983657837,
+      "eval_anatomy_runtime": 0.2675,
+      "eval_anatomy_samples_per_second": 7.476,
+      "eval_anatomy_steps_per_second": 3.738,
+      "step": 16900
+    },
+    {
+      "epoch": 7.032875572201415,
+      "eval_college_mathematics_loss": 2.0515077114105225,
+      "eval_college_mathematics_runtime": 0.2663,
+      "eval_college_mathematics_samples_per_second": 7.51,
+      "eval_college_mathematics_steps_per_second": 3.755,
+      "step": 16900
+    },
+    {
+      "epoch": 7.032875572201415,
+      "eval_international_law_loss": 3.0670435428619385,
+      "eval_international_law_runtime": 0.2682,
+      "eval_international_law_samples_per_second": 7.458,
+      "eval_international_law_steps_per_second": 3.729,
+      "step": 16900
+    },
+    {
+      "epoch": 7.0411985018726595,
+      "grad_norm": 0.31640625,
+      "learning_rate": 1.5060505290409594e-06,
+      "loss": 0.237,
+      "step": 16920
+    },
+    {
+      "epoch": 7.049521431543903,
+      "grad_norm": 0.27734375,
+      "learning_rate": 1.4985583907112188e-06,
+      "loss": 0.2303,
+      "step": 16940
+    },
+    {
+      "epoch": 7.057844361215148,
+      "grad_norm": 0.26953125,
+      "learning_rate": 1.491076950255186e-06,
+      "loss": 0.2309,
+      "step": 16960
+    },
+    {
+      "epoch": 7.066167290886392,
+      "grad_norm": 0.255859375,
+      "learning_rate": 1.4836062875931534e-06,
+      "loss": 0.2307,
+      "step": 16980
+    },
+    {
+      "epoch": 7.074490220557636,
+      "grad_norm": 0.275390625,
+      "learning_rate": 1.4761464825302788e-06,
+      "loss": 0.2316,
+      "step": 17000
+    },
+    {
+      "epoch": 7.074490220557636,
+      "eval_main_loss": 0.2383512556552887,
+      "eval_main_runtime": 6.3399,
+      "eval_main_samples_per_second": 29.969,
+      "eval_main_steps_per_second": 3.786,
+      "step": 17000
+    },
+    {
+      "epoch": 7.074490220557636,
+      "eval_anatomy_loss": 2.8314664363861084,
+      "eval_anatomy_runtime": 0.2674,
+      "eval_anatomy_samples_per_second": 7.481,
+      "eval_anatomy_steps_per_second": 3.74,
+      "step": 17000
+    },
+    {
+      "epoch": 7.074490220557636,
+      "eval_college_mathematics_loss": 2.0510685443878174,
+      "eval_college_mathematics_runtime": 0.2673,
+      "eval_college_mathematics_samples_per_second": 7.482,
+      "eval_college_mathematics_steps_per_second": 3.741,
+      "step": 17000
+    },
+    {
+      "epoch": 7.074490220557636,
+      "eval_international_law_loss": 3.0641367435455322,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.504,
+      "eval_international_law_steps_per_second": 3.752,
+      "step": 17000
+    },
+    {
+      "epoch": 7.082813150228881,
+      "grad_norm": 0.291015625,
+      "learning_rate": 1.4686976147557332e-06,
+      "loss": 0.2291,
+      "step": 17020
+    },
+    {
+      "epoch": 7.091136079900124,
+      "grad_norm": 0.279296875,
+      "learning_rate": 1.461259763841853e-06,
+      "loss": 0.2321,
+      "step": 17040
+    },
+    {
+      "epoch": 7.099459009571369,
+      "grad_norm": 0.291015625,
+      "learning_rate": 1.4538330092432828e-06,
+      "loss": 0.2324,
+      "step": 17060
+    },
+    {
+      "epoch": 7.107781939242614,
+      "grad_norm": 0.294921875,
+      "learning_rate": 1.4464174302961343e-06,
+      "loss": 0.2357,
+      "step": 17080
+    },
+    {
+      "epoch": 7.116104868913857,
+      "grad_norm": 0.28125,
+      "learning_rate": 1.4390131062171378e-06,
+      "loss": 0.2319,
+      "step": 17100
+    },
+    {
+      "epoch": 7.116104868913857,
+      "eval_main_loss": 0.2382993996143341,
+      "eval_main_runtime": 6.329,
+      "eval_main_samples_per_second": 30.021,
+      "eval_main_steps_per_second": 3.792,
+      "step": 17100
+    },
+    {
+      "epoch": 7.116104868913857,
+      "eval_anatomy_loss": 2.833508014678955,
+      "eval_anatomy_runtime": 0.2665,
+      "eval_anatomy_samples_per_second": 7.506,
+      "eval_anatomy_steps_per_second": 3.753,
+      "step": 17100
+    },
+    {
+      "epoch": 7.116104868913857,
+      "eval_college_mathematics_loss": 2.0484731197357178,
+      "eval_college_mathematics_runtime": 0.2653,
+      "eval_college_mathematics_samples_per_second": 7.538,
+      "eval_college_mathematics_steps_per_second": 3.769,
+      "step": 17100
+    },
+    {
+      "epoch": 7.116104868913857,
+      "eval_international_law_loss": 3.0654804706573486,
+      "eval_international_law_runtime": 0.2664,
+      "eval_international_law_samples_per_second": 7.508,
+      "eval_international_law_steps_per_second": 3.754,
+      "step": 17100
+    },
+    {
+      "epoch": 7.124427798585102,
+      "grad_norm": 0.2890625,
+      "learning_rate": 1.4316201161027865e-06,
+      "loss": 0.227,
+      "step": 17120
+    },
+    {
+      "epoch": 7.1327507282563465,
+      "grad_norm": 0.259765625,
+      "learning_rate": 1.4242385389285068e-06,
+      "loss": 0.2312,
+      "step": 17140
+    },
+    {
+      "epoch": 7.14107365792759,
+      "grad_norm": 0.310546875,
+      "learning_rate": 1.416868453547802e-06,
+      "loss": 0.2329,
+      "step": 17160
+    },
+    {
+      "epoch": 7.149396587598835,
+      "grad_norm": 0.255859375,
+      "learning_rate": 1.4095099386914146e-06,
+      "loss": 0.2376,
+      "step": 17180
+    },
+    {
+      "epoch": 7.157719517270079,
+      "grad_norm": 0.294921875,
+      "learning_rate": 1.402163072966488e-06,
+      "loss": 0.2379,
+      "step": 17200
+    },
+    {
+      "epoch": 7.157719517270079,
+      "eval_main_loss": 0.2382867932319641,
+      "eval_main_runtime": 6.3286,
+      "eval_main_samples_per_second": 30.022,
+      "eval_main_steps_per_second": 3.792,
+      "step": 17200
+    },
+    {
+      "epoch": 7.157719517270079,
+      "eval_anatomy_loss": 2.8338112831115723,
+      "eval_anatomy_runtime": 0.267,
+      "eval_anatomy_samples_per_second": 7.491,
+      "eval_anatomy_steps_per_second": 3.746,
+      "step": 17200
+    },
+    {
+      "epoch": 7.157719517270079,
+      "eval_college_mathematics_loss": 2.050143241882324,
+      "eval_college_mathematics_runtime": 0.2655,
+      "eval_college_mathematics_samples_per_second": 7.532,
+      "eval_college_mathematics_steps_per_second": 3.766,
+      "step": 17200
+    },
+    {
+      "epoch": 7.157719517270079,
+      "eval_international_law_loss": 3.067023754119873,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.506,
+      "eval_international_law_steps_per_second": 3.753,
+      "step": 17200
+    },
+    {
+      "epoch": 7.166042446941323,
+      "grad_norm": 0.29296875,
+      "learning_rate": 1.3948279348557225e-06,
+      "loss": 0.234,
+      "step": 17220
+    },
+    {
+      "epoch": 7.174365376612568,
+      "grad_norm": 0.244140625,
+      "learning_rate": 1.3875046027165376e-06,
+      "loss": 0.2318,
+      "step": 17240
+    },
+    {
+      "epoch": 7.182688306283812,
+      "grad_norm": 0.294921875,
+      "learning_rate": 1.380193154780236e-06,
+      "loss": 0.2343,
+      "step": 17260
+    },
+    {
+      "epoch": 7.191011235955056,
+      "grad_norm": 0.2421875,
+      "learning_rate": 1.3728936691511704e-06,
+      "loss": 0.2304,
+      "step": 17280
+    },
+    {
+      "epoch": 7.199334165626301,
+      "grad_norm": 0.271484375,
+      "learning_rate": 1.3656062238059035e-06,
+      "loss": 0.2359,
+      "step": 17300
+    },
+    {
+      "epoch": 7.199334165626301,
+      "eval_main_loss": 0.23832285404205322,
+      "eval_main_runtime": 6.3287,
+      "eval_main_samples_per_second": 30.022,
+      "eval_main_steps_per_second": 3.792,
+      "step": 17300
+    },
+    {
+      "epoch": 7.199334165626301,
+      "eval_anatomy_loss": 2.833261251449585,
+      "eval_anatomy_runtime": 0.2661,
+      "eval_anatomy_samples_per_second": 7.515,
+      "eval_anatomy_steps_per_second": 3.757,
+      "step": 17300
+    },
+    {
+      "epoch": 7.199334165626301,
+      "eval_college_mathematics_loss": 2.0532398223876953,
+      "eval_college_mathematics_runtime": 0.2668,
+      "eval_college_mathematics_samples_per_second": 7.497,
+      "eval_college_mathematics_steps_per_second": 3.748,
+      "step": 17300
+    },
+    {
+      "epoch": 7.199334165626301,
+      "eval_international_law_loss": 3.065138816833496,
+      "eval_international_law_runtime": 0.2663,
+      "eval_international_law_samples_per_second": 7.51,
+      "eval_international_law_steps_per_second": 3.755,
+      "step": 17300
+    },
+    {
+      "epoch": 7.207657095297545,
+      "grad_norm": 0.251953125,
+      "learning_rate": 1.3583308965923778e-06,
+      "loss": 0.2289,
+      "step": 17320
+    },
+    {
+      "epoch": 7.215980024968789,
+      "grad_norm": 0.26171875,
+      "learning_rate": 1.3510677652290882e-06,
+      "loss": 0.2337,
+      "step": 17340
+    },
+    {
+      "epoch": 7.2243029546400335,
+      "grad_norm": 0.2314453125,
+      "learning_rate": 1.3438169073042415e-06,
+      "loss": 0.2328,
+      "step": 17360
+    },
+    {
+      "epoch": 7.232625884311277,
+      "grad_norm": 0.275390625,
+      "learning_rate": 1.3365784002749393e-06,
+      "loss": 0.2327,
+      "step": 17380
+    },
+    {
+      "epoch": 7.240948813982522,
+      "grad_norm": 0.2734375,
+      "learning_rate": 1.3293523214663428e-06,
+      "loss": 0.2336,
+      "step": 17400
+    },
+    {
+      "epoch": 7.240948813982522,
+      "eval_main_loss": 0.23834027349948883,
+      "eval_main_runtime": 6.3246,
+      "eval_main_samples_per_second": 30.041,
+      "eval_main_steps_per_second": 3.795,
+      "step": 17400
+    },
+    {
+      "epoch": 7.240948813982522,
+      "eval_anatomy_loss": 2.8343632221221924,
+      "eval_anatomy_runtime": 0.2656,
+      "eval_anatomy_samples_per_second": 7.53,
+      "eval_anatomy_steps_per_second": 3.765,
+      "step": 17400
+    },
+    {
+      "epoch": 7.240948813982522,
+      "eval_college_mathematics_loss": 2.052730083465576,
+      "eval_college_mathematics_runtime": 0.2666,
+      "eval_college_mathematics_samples_per_second": 7.503,
+      "eval_college_mathematics_steps_per_second": 3.752,
+      "step": 17400
+    },
+    {
+      "epoch": 7.240948813982522,
+      "eval_international_law_loss": 3.066589117050171,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.506,
+      "eval_international_law_steps_per_second": 3.753,
+      "step": 17400
+    },
+    {
+      "epoch": 7.249271743653766,
+      "grad_norm": 0.25390625,
+      "learning_rate": 1.3221387480708475e-06,
+      "loss": 0.2331,
+      "step": 17420
+    },
+    {
+      "epoch": 7.25759467332501,
+      "grad_norm": 0.271484375,
+      "learning_rate": 1.3149377571472655e-06,
+      "loss": 0.2362,
+      "step": 17440
+    },
+    {
+      "epoch": 7.265917602996255,
+      "grad_norm": 0.248046875,
+      "learning_rate": 1.3077494256199892e-06,
+      "loss": 0.2297,
+      "step": 17460
+    },
+    {
+      "epoch": 7.274240532667499,
+      "grad_norm": 0.29296875,
+      "learning_rate": 1.3005738302781839e-06,
+      "loss": 0.2319,
+      "step": 17480
+    },
+    {
+      "epoch": 7.282563462338743,
+      "grad_norm": 0.29296875,
+      "learning_rate": 1.2934110477749584e-06,
+      "loss": 0.2295,
+      "step": 17500
+    },
+    {
+      "epoch": 7.282563462338743,
+      "eval_main_loss": 0.23828759789466858,
+      "eval_main_runtime": 6.3264,
+      "eval_main_samples_per_second": 30.033,
+      "eval_main_steps_per_second": 3.794,
+      "step": 17500
+    },
+    {
+      "epoch": 7.282563462338743,
+      "eval_anatomy_loss": 2.8331925868988037,
+      "eval_anatomy_runtime": 0.2662,
+      "eval_anatomy_samples_per_second": 7.513,
+      "eval_anatomy_steps_per_second": 3.756,
+      "step": 17500
+    },
+    {
+      "epoch": 7.282563462338743,
+      "eval_college_mathematics_loss": 2.0508506298065186,
+      "eval_college_mathematics_runtime": 0.2652,
+      "eval_college_mathematics_samples_per_second": 7.542,
+      "eval_college_mathematics_steps_per_second": 3.771,
+      "step": 17500
+    },
+    {
+      "epoch": 7.282563462338743,
+      "eval_international_law_loss": 3.0639867782592773,
+      "eval_international_law_runtime": 0.2656,
+      "eval_international_law_samples_per_second": 7.532,
+      "eval_international_law_steps_per_second": 3.766,
+      "step": 17500
+    },
+    {
+      "epoch": 7.290886392009988,
+      "grad_norm": 0.263671875,
+      "learning_rate": 1.2862611546265469e-06,
+      "loss": 0.2287,
+      "step": 17520
+    },
+    {
+      "epoch": 7.299209321681232,
+      "grad_norm": 0.27734375,
+      "learning_rate": 1.279124227211498e-06,
+      "loss": 0.2313,
+      "step": 17540
+    },
+    {
+      "epoch": 7.307532251352476,
+      "grad_norm": 0.2890625,
+      "learning_rate": 1.2720003417698506e-06,
+      "loss": 0.2324,
+      "step": 17560
+    },
+    {
+      "epoch": 7.3158551810237205,
+      "grad_norm": 0.298828125,
+      "learning_rate": 1.2648895744023223e-06,
+      "loss": 0.2323,
+      "step": 17580
+    },
+    {
+      "epoch": 7.324178110694964,
+      "grad_norm": 0.318359375,
+      "learning_rate": 1.2577920010695015e-06,
+      "loss": 0.233,
+      "step": 17600
+    },
+    {
+      "epoch": 7.324178110694964,
+      "eval_main_loss": 0.2383326143026352,
+      "eval_main_runtime": 6.331,
+      "eval_main_samples_per_second": 30.011,
+      "eval_main_steps_per_second": 3.791,
+      "step": 17600
+    },
+    {
+      "epoch": 7.324178110694964,
+      "eval_anatomy_loss": 2.833402633666992,
+      "eval_anatomy_runtime": 0.2664,
+      "eval_anatomy_samples_per_second": 7.508,
+      "eval_anatomy_steps_per_second": 3.754,
+      "step": 17600
+    },
+    {
+      "epoch": 7.324178110694964,
+      "eval_college_mathematics_loss": 2.052766799926758,
+      "eval_college_mathematics_runtime": 0.2663,
+      "eval_college_mathematics_samples_per_second": 7.511,
+      "eval_college_mathematics_steps_per_second": 3.755,
+      "step": 17600
+    },
+    {
+      "epoch": 7.324178110694964,
+      "eval_international_law_loss": 3.0673341751098633,
+      "eval_international_law_runtime": 0.2663,
+      "eval_international_law_samples_per_second": 7.511,
+      "eval_international_law_steps_per_second": 3.756,
+      "step": 17600
+    },
+    {
+      "epoch": 7.332501040366209,
+      "grad_norm": 0.279296875,
+      "learning_rate": 1.2507076975910276e-06,
+      "loss": 0.2342,
+      "step": 17620
+    },
+    {
+      "epoch": 7.340823970037453,
+      "grad_norm": 0.322265625,
+      "learning_rate": 1.243636739644787e-06,
+      "loss": 0.2301,
+      "step": 17640
+    },
+    {
+      "epoch": 7.349146899708697,
+      "grad_norm": 0.28515625,
+      "learning_rate": 1.236579202766102e-06,
+      "loss": 0.2286,
+      "step": 17660
+    },
+    {
+      "epoch": 7.357469829379942,
+      "grad_norm": 0.275390625,
+      "learning_rate": 1.229535162346924e-06,
+      "loss": 0.2294,
+      "step": 17680
+    },
+    {
+      "epoch": 7.365792759051186,
+      "grad_norm": 0.35546875,
+      "learning_rate": 1.2225046936350324e-06,
+      "loss": 0.2349,
+      "step": 17700
+    },
+    {
+      "epoch": 7.365792759051186,
+      "eval_main_loss": 0.2383442372083664,
+      "eval_main_runtime": 6.3529,
+      "eval_main_samples_per_second": 29.908,
+      "eval_main_steps_per_second": 3.778,
+      "step": 17700
+    },
+    {
+      "epoch": 7.365792759051186,
+      "eval_anatomy_loss": 2.8332571983337402,
+      "eval_anatomy_runtime": 0.2669,
+      "eval_anatomy_samples_per_second": 7.493,
+      "eval_anatomy_steps_per_second": 3.746,
+      "step": 17700
+    },
+    {
+      "epoch": 7.365792759051186,
+      "eval_college_mathematics_loss": 2.0507707595825195,
+      "eval_college_mathematics_runtime": 0.267,
+      "eval_college_mathematics_samples_per_second": 7.492,
+      "eval_college_mathematics_steps_per_second": 3.746,
+      "step": 17700
+    },
+    {
+      "epoch": 7.365792759051186,
+      "eval_international_law_loss": 3.0661237239837646,
+      "eval_international_law_runtime": 0.2671,
+      "eval_international_law_samples_per_second": 7.488,
+      "eval_international_law_steps_per_second": 3.744,
+      "step": 17700
+    },
+    {
+      "epoch": 7.37411568872243,
+      "grad_norm": 0.2890625,
+      "learning_rate": 1.2154878717332235e-06,
+      "loss": 0.2334,
+      "step": 17720
+    },
+    {
+      "epoch": 7.3824386183936745,
+      "grad_norm": 0.25,
+      "learning_rate": 1.208484771598513e-06,
+      "loss": 0.2292,
+      "step": 17740
+    },
+    {
+      "epoch": 7.390761548064919,
+      "grad_norm": 0.29296875,
+      "learning_rate": 1.2014954680413334e-06,
+      "loss": 0.2273,
+      "step": 17760
+    },
+    {
+      "epoch": 7.399084477736163,
+      "grad_norm": 0.279296875,
+      "learning_rate": 1.1945200357247386e-06,
+      "loss": 0.2305,
+      "step": 17780
+    },
+    {
+      "epoch": 7.407407407407407,
+      "grad_norm": 0.271484375,
+      "learning_rate": 1.1875585491636e-06,
+      "loss": 0.2317,
+      "step": 17800
+    },
+    {
+      "epoch": 7.407407407407407,
+      "eval_main_loss": 0.23827992379665375,
+      "eval_main_runtime": 6.3476,
+      "eval_main_samples_per_second": 29.932,
+      "eval_main_steps_per_second": 3.781,
+      "step": 17800
+    },
+    {
+      "epoch": 7.407407407407407,
+      "eval_anatomy_loss": 2.832733631134033,
+      "eval_anatomy_runtime": 0.2672,
+      "eval_anatomy_samples_per_second": 7.484,
+      "eval_anatomy_steps_per_second": 3.742,
+      "step": 17800
+    },
+    {
+      "epoch": 7.407407407407407,
+      "eval_college_mathematics_loss": 2.052572250366211,
+      "eval_college_mathematics_runtime": 0.2671,
+      "eval_college_mathematics_samples_per_second": 7.489,
+      "eval_college_mathematics_steps_per_second": 3.744,
+      "step": 17800
+    },
+    {
+      "epoch": 7.407407407407407,
+      "eval_international_law_loss": 3.064929723739624,
+      "eval_international_law_runtime": 0.2672,
+      "eval_international_law_samples_per_second": 7.484,
+      "eval_international_law_steps_per_second": 3.742,
+      "step": 17800
+    },
+    {
+      "epoch": 7.415730337078652,
+      "grad_norm": 0.27734375,
+      "learning_rate": 1.180611082723814e-06,
+      "loss": 0.2316,
+      "step": 17820
+    },
+    {
+      "epoch": 7.424053266749896,
+      "grad_norm": 0.28515625,
+      "learning_rate": 1.1736777106215118e-06,
+      "loss": 0.2336,
+      "step": 17840
+    },
+    {
+      "epoch": 7.43237619642114,
+      "grad_norm": 0.255859375,
+      "learning_rate": 1.1667585069222554e-06,
+      "loss": 0.2324,
+      "step": 17860
+    },
+    {
+      "epoch": 7.440699126092385,
+      "grad_norm": 0.271484375,
+      "learning_rate": 1.1598535455402584e-06,
+      "loss": 0.2301,
+      "step": 17880
+    },
+    {
+      "epoch": 7.449022055763629,
+      "grad_norm": 0.30859375,
+      "learning_rate": 1.15296290023759e-06,
+      "loss": 0.2282,
+      "step": 17900
+    },
+    {
+      "epoch": 7.449022055763629,
+      "eval_main_loss": 0.2383432686328888,
+      "eval_main_runtime": 6.356,
+      "eval_main_samples_per_second": 29.893,
+      "eval_main_steps_per_second": 3.776,
+      "step": 17900
+    },
+    {
+      "epoch": 7.449022055763629,
+      "eval_anatomy_loss": 2.833138942718506,
+      "eval_anatomy_runtime": 0.2678,
+      "eval_anatomy_samples_per_second": 7.468,
+      "eval_anatomy_steps_per_second": 3.734,
+      "step": 17900
+    },
+    {
+      "epoch": 7.449022055763629,
+      "eval_college_mathematics_loss": 2.052255868911743,
+      "eval_college_mathematics_runtime": 0.2683,
+      "eval_college_mathematics_samples_per_second": 7.454,
+      "eval_college_mathematics_steps_per_second": 3.727,
+      "step": 17900
+    },
+    {
+      "epoch": 7.449022055763629,
+      "eval_international_law_loss": 3.065342664718628,
+      "eval_international_law_runtime": 0.266,
+      "eval_international_law_samples_per_second": 7.519,
+      "eval_international_law_steps_per_second": 3.76,
+      "step": 17900
+    },
+    {
+      "epoch": 7.457344985434873,
+      "grad_norm": 0.283203125,
+      "learning_rate": 1.1460866446233857e-06,
+      "loss": 0.2326,
+      "step": 17920
+    },
+    {
+      "epoch": 7.465667915106117,
+      "grad_norm": 0.26953125,
+      "learning_rate": 1.1392248521530692e-06,
+      "loss": 0.2321,
+      "step": 17940
+    },
+    {
+      "epoch": 7.4739908447773615,
+      "grad_norm": 0.3125,
+      "learning_rate": 1.132377596127554e-06,
+      "loss": 0.2294,
+      "step": 17960
+    },
+    {
+      "epoch": 7.482313774448606,
+      "grad_norm": 0.283203125,
+      "learning_rate": 1.1255449496924767e-06,
+      "loss": 0.2333,
+      "step": 17980
+    },
+    {
+      "epoch": 7.49063670411985,
+      "grad_norm": 0.267578125,
+      "learning_rate": 1.1187269858374017e-06,
+      "loss": 0.2293,
+      "step": 18000
+    },
+    {
+      "epoch": 7.49063670411985,
+      "eval_main_loss": 0.2383362054824829,
+      "eval_main_runtime": 6.3387,
+      "eval_main_samples_per_second": 29.974,
+      "eval_main_steps_per_second": 3.786,
+      "step": 18000
+    },
+    {
+      "epoch": 7.49063670411985,
+      "eval_anatomy_loss": 2.832285165786743,
+      "eval_anatomy_runtime": 0.267,
+      "eval_anatomy_samples_per_second": 7.492,
+      "eval_anatomy_steps_per_second": 3.746,
+      "step": 18000
+    },
+    {
+      "epoch": 7.49063670411985,
+      "eval_college_mathematics_loss": 2.0504801273345947,
+      "eval_college_mathematics_runtime": 0.2653,
+      "eval_college_mathematics_samples_per_second": 7.54,
+      "eval_college_mathematics_steps_per_second": 3.77,
+      "step": 18000
+    },
+    {
+      "epoch": 7.49063670411985,
+      "eval_international_law_loss": 3.0625662803649902,
+      "eval_international_law_runtime": 0.2658,
+      "eval_international_law_samples_per_second": 7.524,
+      "eval_international_law_steps_per_second": 3.762,
+      "step": 18000
+    },
+    {
+      "epoch": 7.498959633791094,
+      "grad_norm": 0.279296875,
+      "learning_rate": 1.1119237773950484e-06,
+      "loss": 0.2286,
+      "step": 18020
+    },
+    {
+      "epoch": 7.507282563462339,
+      "grad_norm": 0.2431640625,
+      "learning_rate": 1.1051353970405142e-06,
+      "loss": 0.2315,
+      "step": 18040
+    },
+    {
+      "epoch": 7.515605493133583,
+      "grad_norm": 0.341796875,
+      "learning_rate": 1.0983619172904935e-06,
+      "loss": 0.2284,
+      "step": 18060
+    },
+    {
+      "epoch": 7.523928422804827,
+      "grad_norm": 0.3046875,
+      "learning_rate": 1.0916034105025052e-06,
+      "loss": 0.2282,
+      "step": 18080
+    },
+    {
+      "epoch": 7.532251352476072,
+      "grad_norm": 0.248046875,
+      "learning_rate": 1.0848599488741208e-06,
+      "loss": 0.2277,
+      "step": 18100
+    },
+    {
+      "epoch": 7.532251352476072,
+      "eval_main_loss": 0.2383170872926712,
+      "eval_main_runtime": 6.3286,
+      "eval_main_samples_per_second": 30.023,
+      "eval_main_steps_per_second": 3.792,
+      "step": 18100
+    },
+    {
+      "epoch": 7.532251352476072,
+      "eval_anatomy_loss": 2.834378480911255,
+      "eval_anatomy_runtime": 0.2677,
+      "eval_anatomy_samples_per_second": 7.472,
+      "eval_anatomy_steps_per_second": 3.736,
+      "step": 18100
+    },
+    {
+      "epoch": 7.532251352476072,
+      "eval_college_mathematics_loss": 2.0512096881866455,
+      "eval_college_mathematics_runtime": 0.2666,
+      "eval_college_mathematics_samples_per_second": 7.501,
+      "eval_college_mathematics_steps_per_second": 3.75,
+      "step": 18100
+    },
+    {
+      "epoch": 7.532251352476072,
+      "eval_international_law_loss": 3.066354513168335,
+      "eval_international_law_runtime": 0.2663,
+      "eval_international_law_samples_per_second": 7.51,
+      "eval_international_law_steps_per_second": 3.755,
+      "step": 18100
+    },
+    {
+      "epoch": 7.540574282147316,
+      "grad_norm": 0.3203125,
+      "learning_rate": 1.078131604442193e-06,
+      "loss": 0.2295,
+      "step": 18120
+    },
+    {
+      "epoch": 7.54889721181856,
+      "grad_norm": 0.310546875,
+      "learning_rate": 1.0714184490820842e-06,
+      "loss": 0.2296,
+      "step": 18140
+    },
+    {
+      "epoch": 7.557220141489804,
+      "grad_norm": 0.2890625,
+      "learning_rate": 1.0647205545068992e-06,
+      "loss": 0.2311,
+      "step": 18160
+    },
+    {
+      "epoch": 7.5655430711610485,
+      "grad_norm": 0.259765625,
+      "learning_rate": 1.0580379922667241e-06,
+      "loss": 0.2333,
+      "step": 18180
+    },
+    {
+      "epoch": 7.573866000832293,
+      "grad_norm": 0.3046875,
+      "learning_rate": 1.0513708337478509e-06,
+      "loss": 0.2304,
+      "step": 18200
+    },
+    {
+      "epoch": 7.573866000832293,
+      "eval_main_loss": 0.2383020520210266,
+      "eval_main_runtime": 6.3287,
+      "eval_main_samples_per_second": 30.022,
+      "eval_main_steps_per_second": 3.792,
+      "step": 18200
+    },
+    {
+      "epoch": 7.573866000832293,
+      "eval_anatomy_loss": 2.834932804107666,
+      "eval_anatomy_runtime": 0.2668,
+      "eval_anatomy_samples_per_second": 7.496,
+      "eval_anatomy_steps_per_second": 3.748,
+      "step": 18200
+    },
+    {
+      "epoch": 7.573866000832293,
+      "eval_college_mathematics_loss": 2.0513229370117188,
+      "eval_college_mathematics_runtime": 0.2658,
+      "eval_college_mathematics_samples_per_second": 7.525,
+      "eval_college_mathematics_steps_per_second": 3.762,
+      "step": 18200
+    },
+    {
+      "epoch": 7.573866000832293,
+      "eval_international_law_loss": 3.0660512447357178,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.504,
+      "eval_international_law_steps_per_second": 3.752,
+      "step": 18200
+    },
+    {
+      "epoch": 7.582188930503538,
+      "grad_norm": 0.279296875,
+      "learning_rate": 1.044719150172028e-06,
+      "loss": 0.2259,
+      "step": 18220
+    },
+    {
+      "epoch": 7.590511860174781,
+      "grad_norm": 0.3125,
+      "learning_rate": 1.0380830125956897e-06,
+      "loss": 0.2273,
+      "step": 18240
+    },
+    {
+      "epoch": 7.598834789846026,
+      "grad_norm": 0.3515625,
+      "learning_rate": 1.0314624919092011e-06,
+      "loss": 0.2365,
+      "step": 18260
+    },
+    {
+      "epoch": 7.60715771951727,
+      "grad_norm": 0.322265625,
+      "learning_rate": 1.024857658836102e-06,
+      "loss": 0.2327,
+      "step": 18280
+    },
+    {
+      "epoch": 7.615480649188514,
+      "grad_norm": 0.294921875,
+      "learning_rate": 1.0182685839323475e-06,
+      "loss": 0.232,
+      "step": 18300
+    },
+    {
+      "epoch": 7.615480649188514,
+      "eval_main_loss": 0.23833827674388885,
+      "eval_main_runtime": 6.3315,
+      "eval_main_samples_per_second": 30.009,
+      "eval_main_steps_per_second": 3.791,
+      "step": 18300
+    },
+    {
+      "epoch": 7.615480649188514,
+      "eval_anatomy_loss": 2.8317782878875732,
+      "eval_anatomy_runtime": 0.2681,
+      "eval_anatomy_samples_per_second": 7.46,
+      "eval_anatomy_steps_per_second": 3.73,
+      "step": 18300
+    },
+    {
+      "epoch": 7.615480649188514,
+      "eval_college_mathematics_loss": 2.0530128479003906,
+      "eval_college_mathematics_runtime": 0.2671,
+      "eval_college_mathematics_samples_per_second": 7.489,
+      "eval_college_mathematics_steps_per_second": 3.744,
+      "step": 18300
+    },
+    {
+      "epoch": 7.615480649188514,
+      "eval_international_law_loss": 3.064087152481079,
+      "eval_international_law_runtime": 0.2649,
+      "eval_international_law_samples_per_second": 7.551,
+      "eval_international_law_steps_per_second": 3.776,
+      "step": 18300
+    },
+    {
+      "epoch": 7.623803578859759,
+      "grad_norm": 0.2734375,
+      "learning_rate": 1.0116953375855565e-06,
+      "loss": 0.2331,
+      "step": 18320
+    },
+    {
+      "epoch": 7.632126508531003,
+      "grad_norm": 0.2138671875,
+      "learning_rate": 1.0051379900142635e-06,
+      "loss": 0.2336,
+      "step": 18340
+    },
+    {
+      "epoch": 7.640449438202247,
+      "grad_norm": 0.298828125,
+      "learning_rate": 9.98596611267158e-07,
+      "loss": 0.2308,
+      "step": 18360
+    },
+    {
+      "epoch": 7.648772367873492,
+      "grad_norm": 0.267578125,
+      "learning_rate": 9.920712712223494e-07,
+      "loss": 0.2258,
+      "step": 18380
+    },
+    {
+      "epoch": 7.6570952975447355,
+      "grad_norm": 0.28125,
+      "learning_rate": 9.855620395866107e-07,
+      "loss": 0.2336,
+      "step": 18400
+    },
+    {
+      "epoch": 7.6570952975447355,
+      "eval_main_loss": 0.23823559284210205,
+      "eval_main_runtime": 6.3201,
+      "eval_main_samples_per_second": 30.063,
+      "eval_main_steps_per_second": 3.797,
+      "step": 18400
+    },
+    {
+      "epoch": 7.6570952975447355,
+      "eval_anatomy_loss": 2.834606409072876,
+      "eval_anatomy_runtime": 0.2664,
+      "eval_anatomy_samples_per_second": 7.508,
+      "eval_anatomy_steps_per_second": 3.754,
+      "step": 18400
+    },
+    {
+      "epoch": 7.6570952975447355,
+      "eval_college_mathematics_loss": 2.0500612258911133,
+      "eval_college_mathematics_runtime": 0.2665,
+      "eval_college_mathematics_samples_per_second": 7.506,
+      "eval_college_mathematics_steps_per_second": 3.753,
+      "step": 18400
+    },
+    {
+      "epoch": 7.6570952975447355,
+      "eval_international_law_loss": 3.0666239261627197,
+      "eval_international_law_runtime": 0.2661,
+      "eval_international_law_samples_per_second": 7.515,
+      "eval_international_law_steps_per_second": 3.757,
+      "step": 18400
+    },
+    {
+      "epoch": 7.66541822721598,
+      "grad_norm": 0.26953125,
+      "learning_rate": 9.790689858946374e-07,
+      "loss": 0.2301,
+      "step": 18420
+    },
+    {
+      "epoch": 7.673741156887225,
+      "grad_norm": 0.2490234375,
+      "learning_rate": 9.725921795083063e-07,
+      "loss": 0.2327,
+      "step": 18440
+    },
+    {
+      "epoch": 7.682064086558468,
+      "grad_norm": 0.2578125,
+      "learning_rate": 9.661316896159313e-07,
+      "loss": 0.2304,
+      "step": 18460
+    },
+    {
+      "epoch": 7.690387016229713,
+      "grad_norm": 0.263671875,
+      "learning_rate": 9.59687585231526e-07,
+      "loss": 0.2282,
+      "step": 18480
+    },
+    {
+      "epoch": 7.698709945900957,
+      "grad_norm": 0.287109375,
+      "learning_rate": 9.53259935194066e-07,
+      "loss": 0.2313,
+      "step": 18500
+    },
+    {
+      "epoch": 7.698709945900957,
+      "eval_main_loss": 0.23836714029312134,
+      "eval_main_runtime": 6.351,
+      "eval_main_samples_per_second": 29.917,
+      "eval_main_steps_per_second": 3.779,
+      "step": 18500
+    },
+    {
+      "epoch": 7.698709945900957,
+      "eval_anatomy_loss": 2.8336453437805176,
+      "eval_anatomy_runtime": 0.2666,
+      "eval_anatomy_samples_per_second": 7.502,
+      "eval_anatomy_steps_per_second": 3.751,
+      "step": 18500
+    },
+    {
+      "epoch": 7.698709945900957,
+      "eval_college_mathematics_loss": 2.0492594242095947,
+      "eval_college_mathematics_runtime": 0.2688,
+      "eval_college_mathematics_samples_per_second": 7.442,
+      "eval_college_mathematics_steps_per_second": 3.721,
+      "step": 18500
+    },
+    {
+      "epoch": 7.698709945900957,
+      "eval_international_law_loss": 3.0663108825683594,
+      "eval_international_law_runtime": 0.267,
+      "eval_international_law_samples_per_second": 7.489,
+      "eval_international_law_steps_per_second": 3.745,
+      "step": 18500
+    },
+    {
+      "epoch": 7.707032875572201,
+      "grad_norm": 0.302734375,
+      "learning_rate": 9.468488081667537e-07,
+      "loss": 0.2357,
+      "step": 18520
+    },
+    {
+      "epoch": 7.715355805243446,
+      "grad_norm": 0.3515625,
+      "learning_rate": 9.404542726362872e-07,
+      "loss": 0.231,
+      "step": 18540
+    },
+    {
+      "epoch": 7.72367873491469,
+      "grad_norm": 0.30078125,
+      "learning_rate": 9.34076396912123e-07,
+      "loss": 0.234,
+      "step": 18560
+    },
+    {
+      "epoch": 7.732001664585934,
+      "grad_norm": 0.2412109375,
+      "learning_rate": 9.277152491257515e-07,
+      "loss": 0.2341,
+      "step": 18580
+    },
+    {
+      "epoch": 7.740324594257179,
+      "grad_norm": 0.296875,
+      "learning_rate": 9.213708972299662e-07,
+      "loss": 0.2339,
+      "step": 18600
+    },
+    {
+      "epoch": 7.740324594257179,
+      "eval_main_loss": 0.23831741511821747,
+      "eval_main_runtime": 6.3514,
+      "eval_main_samples_per_second": 29.915,
+      "eval_main_steps_per_second": 3.779,
+      "step": 18600
+    },
+    {
+      "epoch": 7.740324594257179,
+      "eval_anatomy_loss": 2.8322861194610596,
+      "eval_anatomy_runtime": 0.2674,
+      "eval_anatomy_samples_per_second": 7.479,
+      "eval_anatomy_steps_per_second": 3.74,
+      "step": 18600
+    },
+    {
+      "epoch": 7.740324594257179,
+      "eval_college_mathematics_loss": 2.050147771835327,
+      "eval_college_mathematics_runtime": 0.267,
+      "eval_college_mathematics_samples_per_second": 7.49,
+      "eval_college_mathematics_steps_per_second": 3.745,
+      "step": 18600
+    },
+    {
+      "epoch": 7.740324594257179,
+      "eval_international_law_loss": 3.066422700881958,
+      "eval_international_law_runtime": 0.2662,
+      "eval_international_law_samples_per_second": 7.512,
+      "eval_international_law_steps_per_second": 3.756,
+      "step": 18600
+    },
+    {
+      "epoch": 7.7486475239284225,
+      "grad_norm": 0.271484375,
+      "learning_rate": 9.150434089981413e-07,
+      "loss": 0.234,
+      "step": 18620
+    },
+    {
+      "epoch": 7.756970453599667,
+      "grad_norm": 0.2578125,
+      "learning_rate": 9.087328520235028e-07,
+      "loss": 0.2348,
+      "step": 18640
+    },
+    {
+      "epoch": 7.765293383270912,
+      "grad_norm": 0.275390625,
+      "learning_rate": 9.024392937184096e-07,
+      "loss": 0.2315,
+      "step": 18660
+    },
+    {
+      "epoch": 7.773616312942155,
+      "grad_norm": 0.26953125,
+      "learning_rate": 8.961628013136351e-07,
+      "loss": 0.2295,
+      "step": 18680
+    },
+    {
+      "epoch": 7.7819392426134,
+      "grad_norm": 0.259765625,
+      "learning_rate": 8.899034418576413e-07,
+      "loss": 0.2329,
+      "step": 18700
+    },
+    {
+      "epoch": 7.7819392426134,
+      "eval_main_loss": 0.2382706105709076,
+      "eval_main_runtime": 6.3343,
+      "eval_main_samples_per_second": 29.996,
+      "eval_main_steps_per_second": 3.789,
+      "step": 18700
+    },
+    {
+      "epoch": 7.7819392426134,
+      "eval_anatomy_loss": 2.832164764404297,
+      "eval_anatomy_runtime": 0.2662,
+      "eval_anatomy_samples_per_second": 7.514,
+      "eval_anatomy_steps_per_second": 3.757,
+      "step": 18700
+    },
+    {
+      "epoch": 7.7819392426134,
+      "eval_college_mathematics_loss": 2.0492961406707764,
+      "eval_college_mathematics_runtime": 0.267,
+      "eval_college_mathematics_samples_per_second": 7.491,
+      "eval_college_mathematics_steps_per_second": 3.745,
+      "step": 18700
+    },
+    {
+      "epoch": 7.7819392426134,
+      "eval_international_law_loss": 3.065026044845581,
+      "eval_international_law_runtime": 0.2654,
+      "eval_international_law_samples_per_second": 7.537,
+      "eval_international_law_steps_per_second": 3.768,
+      "step": 18700
+    },
+    {
+      "epoch": 7.790262172284645,
+      "grad_norm": 0.25,
+      "learning_rate": 8.836612822158743e-07,
+      "loss": 0.2279,
+      "step": 18720
+    },
+    {
+      "epoch": 7.798585101955888,
+      "grad_norm": 0.318359375,
+      "learning_rate": 8.774363890700394e-07,
+      "loss": 0.2337,
+      "step": 18740
+    },
+    {
+      "epoch": 7.806908031627133,
+      "grad_norm": 0.234375,
+      "learning_rate": 8.712288289173937e-07,
+      "loss": 0.2348,
+      "step": 18760
+    },
+    {
+      "epoch": 7.8152309612983775,
+      "grad_norm": 0.263671875,
+      "learning_rate": 8.650386680700373e-07,
+      "loss": 0.2285,
+      "step": 18780
+    },
+    {
+      "epoch": 7.823553890969621,
+      "grad_norm": 0.26171875,
+      "learning_rate": 8.588659726541998e-07,
+      "loss": 0.2324,
+      "step": 18800
+    },
+    {
+      "epoch": 7.823553890969621,
+      "eval_main_loss": 0.23833107948303223,
+      "eval_main_runtime": 6.3262,
+      "eval_main_samples_per_second": 30.034,
+      "eval_main_steps_per_second": 3.794,
+      "step": 18800
+    },
+    {
+      "epoch": 7.823553890969621,
+      "eval_anatomy_loss": 2.8310635089874268,
+      "eval_anatomy_runtime": 0.2659,
+      "eval_anatomy_samples_per_second": 7.522,
+      "eval_anatomy_steps_per_second": 3.761,
+      "step": 18800
+    },
+    {
+      "epoch": 7.823553890969621,
+      "eval_college_mathematics_loss": 2.0509285926818848,
+      "eval_college_mathematics_runtime": 0.2668,
+      "eval_college_mathematics_samples_per_second": 7.496,
+      "eval_college_mathematics_steps_per_second": 3.748,
+      "step": 18800
+    },
+    {
+      "epoch": 7.823553890969621,
+      "eval_international_law_loss": 3.065098285675049,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.506,
+      "eval_international_law_steps_per_second": 3.753,
+      "step": 18800
+    },
+    {
+      "epoch": 7.831876820640866,
+      "grad_norm": 0.306640625,
+      "learning_rate": 8.527108086095375e-07,
+      "loss": 0.232,
+      "step": 18820
+    },
+    {
+      "epoch": 7.8401997503121095,
+      "grad_norm": 0.2890625,
+      "learning_rate": 8.465732416884312e-07,
+      "loss": 0.2316,
+      "step": 18840
+    },
+    {
+      "epoch": 7.848522679983354,
+      "grad_norm": 0.259765625,
+      "learning_rate": 8.404533374552751e-07,
+      "loss": 0.233,
+      "step": 18860
+    },
+    {
+      "epoch": 7.856845609654599,
+      "grad_norm": 0.314453125,
+      "learning_rate": 8.343511612857879e-07,
+      "loss": 0.2333,
+      "step": 18880
+    },
+    {
+      "epoch": 7.865168539325842,
+      "grad_norm": 0.298828125,
+      "learning_rate": 8.282667783663056e-07,
+      "loss": 0.235,
+      "step": 18900
+    },
+    {
+      "epoch": 7.865168539325842,
+      "eval_main_loss": 0.23828402161598206,
+      "eval_main_runtime": 6.3357,
+      "eval_main_samples_per_second": 29.989,
+      "eval_main_steps_per_second": 3.788,
+      "step": 18900
+    },
+    {
+      "epoch": 7.865168539325842,
+      "eval_anatomy_loss": 2.832879066467285,
+      "eval_anatomy_runtime": 0.2655,
+      "eval_anatomy_samples_per_second": 7.532,
+      "eval_anatomy_steps_per_second": 3.766,
+      "step": 18900
+    },
+    {
+      "epoch": 7.865168539325842,
+      "eval_college_mathematics_loss": 2.05184006690979,
+      "eval_college_mathematics_runtime": 0.2659,
+      "eval_college_mathematics_samples_per_second": 7.521,
+      "eval_college_mathematics_steps_per_second": 3.761,
+      "step": 18900
+    },
+    {
+      "epoch": 7.865168539325842,
+      "eval_international_law_loss": 3.065366506576538,
+      "eval_international_law_runtime": 0.2648,
+      "eval_international_law_samples_per_second": 7.553,
+      "eval_international_law_steps_per_second": 3.776,
+      "step": 18900
+    },
+    {
+      "epoch": 7.873491468997087,
+      "grad_norm": 0.2890625,
+      "learning_rate": 8.222002536930887e-07,
+      "loss": 0.2314,
+      "step": 18920
+    },
+    {
+      "epoch": 7.881814398668332,
+      "grad_norm": 0.26171875,
+      "learning_rate": 8.161516520716287e-07,
+      "loss": 0.2328,
+      "step": 18940
+    },
+    {
+      "epoch": 7.890137328339575,
+      "grad_norm": 0.279296875,
+      "learning_rate": 8.101210381159533e-07,
+      "loss": 0.229,
+      "step": 18960
+    },
+    {
+      "epoch": 7.89846025801082,
+      "grad_norm": 0.267578125,
+      "learning_rate": 8.041084762479376e-07,
+      "loss": 0.2307,
+      "step": 18980
+    },
+    {
+      "epoch": 7.9067831876820645,
+      "grad_norm": 0.291015625,
+      "learning_rate": 7.981140306966154e-07,
+      "loss": 0.2328,
+      "step": 19000
+    },
+    {
+      "epoch": 7.9067831876820645,
+      "eval_main_loss": 0.23827558755874634,
+      "eval_main_runtime": 6.3308,
+      "eval_main_samples_per_second": 30.012,
+      "eval_main_steps_per_second": 3.791,
+      "step": 19000
+    },
+    {
+      "epoch": 7.9067831876820645,
+      "eval_anatomy_loss": 2.8324079513549805,
+      "eval_anatomy_runtime": 0.2667,
+      "eval_anatomy_samples_per_second": 7.5,
+      "eval_anatomy_steps_per_second": 3.75,
+      "step": 19000
+    },
+    {
+      "epoch": 7.9067831876820645,
+      "eval_college_mathematics_loss": 2.050251007080078,
+      "eval_college_mathematics_runtime": 0.266,
+      "eval_college_mathematics_samples_per_second": 7.519,
+      "eval_college_mathematics_steps_per_second": 3.76,
+      "step": 19000
+    },
+    {
+      "epoch": 7.9067831876820645,
+      "eval_international_law_loss": 3.0664069652557373,
+      "eval_international_law_runtime": 0.2674,
+      "eval_international_law_samples_per_second": 7.478,
+      "eval_international_law_steps_per_second": 3.739,
+      "step": 19000
+    },
+    {
+      "epoch": 7.915106117353308,
+      "grad_norm": 0.306640625,
+      "learning_rate": 7.921377654974955e-07,
+      "loss": 0.2349,
+      "step": 19020
+    },
+    {
+      "epoch": 7.923429047024553,
+      "grad_norm": 0.30859375,
+      "learning_rate": 7.861797444918731e-07,
+      "loss": 0.2338,
+      "step": 19040
+    },
+    {
+      "epoch": 7.9317519766957965,
+      "grad_norm": 0.265625,
+      "learning_rate": 7.802400313261505e-07,
+      "loss": 0.2378,
+      "step": 19060
+    },
+    {
+      "epoch": 7.940074906367041,
+      "grad_norm": 0.302734375,
+      "learning_rate": 7.743186894511603e-07,
+      "loss": 0.2277,
+      "step": 19080
+    },
+    {
+      "epoch": 7.948397836038286,
+      "grad_norm": 0.318359375,
+      "learning_rate": 7.684157821214783e-07,
+      "loss": 0.2305,
+      "step": 19100
+    },
+    {
+      "epoch": 7.948397836038286,
+      "eval_main_loss": 0.23835448920726776,
+      "eval_main_runtime": 6.3314,
+      "eval_main_samples_per_second": 30.009,
+      "eval_main_steps_per_second": 3.791,
+      "step": 19100
+    },
+    {
+      "epoch": 7.948397836038286,
+      "eval_anatomy_loss": 2.8330698013305664,
+      "eval_anatomy_runtime": 0.2647,
+      "eval_anatomy_samples_per_second": 7.555,
+      "eval_anatomy_steps_per_second": 3.777,
+      "step": 19100
+    },
+    {
+      "epoch": 7.948397836038286,
+      "eval_college_mathematics_loss": 2.0542187690734863,
+      "eval_college_mathematics_runtime": 0.2653,
+      "eval_college_mathematics_samples_per_second": 7.537,
+      "eval_college_mathematics_steps_per_second": 3.769,
+      "step": 19100
+    },
+    {
+      "epoch": 7.948397836038286,
+      "eval_international_law_loss": 3.0674057006835938,
+      "eval_international_law_runtime": 0.2667,
+      "eval_international_law_samples_per_second": 7.499,
+      "eval_international_law_steps_per_second": 3.749,
+      "step": 19100
+    },
+    {
+      "epoch": 7.95672076570953,
+      "grad_norm": 0.2060546875,
+      "learning_rate": 7.625313723947592e-07,
+      "loss": 0.2358,
+      "step": 19120
+    },
+    {
+      "epoch": 7.965043695380774,
+      "grad_norm": 0.279296875,
+      "learning_rate": 7.566655231310551e-07,
+      "loss": 0.2339,
+      "step": 19140
+    },
+    {
+      "epoch": 7.9733666250520185,
+      "grad_norm": 0.26171875,
+      "learning_rate": 7.508182969921463e-07,
+      "loss": 0.2359,
+      "step": 19160
+    },
+    {
+      "epoch": 7.981689554723262,
+      "grad_norm": 0.333984375,
+      "learning_rate": 7.449897564408743e-07,
+      "loss": 0.2312,
+      "step": 19180
+    },
+    {
+      "epoch": 7.990012484394507,
+      "grad_norm": 0.287109375,
+      "learning_rate": 7.391799637404675e-07,
+      "loss": 0.2328,
+      "step": 19200
+    },
+    {
+      "epoch": 7.990012484394507,
+      "eval_main_loss": 0.238310769200325,
+      "eval_main_runtime": 6.3251,
+      "eval_main_samples_per_second": 30.039,
+      "eval_main_steps_per_second": 3.794,
+      "step": 19200
+    },
+    {
+      "epoch": 7.990012484394507,
+      "eval_anatomy_loss": 2.834007501602173,
+      "eval_anatomy_runtime": 0.266,
+      "eval_anatomy_samples_per_second": 7.518,
+      "eval_anatomy_steps_per_second": 3.759,
+      "step": 19200
+    },
+    {
+      "epoch": 7.990012484394507,
+      "eval_college_mathematics_loss": 2.0495834350585938,
+      "eval_college_mathematics_runtime": 0.2657,
+      "eval_college_mathematics_samples_per_second": 7.528,
+      "eval_college_mathematics_steps_per_second": 3.764,
+      "step": 19200
+    },
+    {
+      "epoch": 7.990012484394507,
+      "eval_international_law_loss": 3.065659523010254,
+      "eval_international_law_runtime": 0.2672,
+      "eval_international_law_samples_per_second": 7.484,
+      "eval_international_law_steps_per_second": 3.742,
+      "step": 19200
+    },
+    {
+      "epoch": 7.998335414065751,
+      "grad_norm": 0.232421875,
+      "learning_rate": 7.333889809538869e-07,
+      "loss": 0.2292,
+      "step": 19220
+    },
+    {
+      "epoch": 8.006658343736996,
+      "grad_norm": 0.263671875,
+      "learning_rate": 7.276168699431527e-07,
+      "loss": 0.2287,
+      "step": 19240
+    },
+    {
+      "epoch": 8.014981273408239,
+      "grad_norm": 0.333984375,
+      "learning_rate": 7.218636923686889e-07,
+      "loss": 0.2308,
+      "step": 19260
+    },
+    {
+      "epoch": 8.023304203079483,
+      "grad_norm": 0.3046875,
+      "learning_rate": 7.16129509688665e-07,
+      "loss": 0.2346,
+      "step": 19280
+    },
+    {
+      "epoch": 8.031627132750728,
+      "grad_norm": 0.263671875,
+      "learning_rate": 7.104143831583368e-07,
+      "loss": 0.2292,
+      "step": 19300
+    },
+    {
+      "epoch": 8.031627132750728,
+      "eval_main_loss": 0.2382253110408783,
+      "eval_main_runtime": 6.3318,
+      "eval_main_samples_per_second": 30.007,
+      "eval_main_steps_per_second": 3.79,
+      "step": 19300
+    },
+    {
+      "epoch": 8.031627132750728,
+      "eval_anatomy_loss": 2.833739757537842,
+      "eval_anatomy_runtime": 0.2665,
+      "eval_anatomy_samples_per_second": 7.505,
+      "eval_anatomy_steps_per_second": 3.752,
+      "step": 19300
+    },
+    {
+      "epoch": 8.031627132750728,
+      "eval_college_mathematics_loss": 2.053969144821167,
+      "eval_college_mathematics_runtime": 0.2672,
+      "eval_college_mathematics_samples_per_second": 7.484,
+      "eval_college_mathematics_steps_per_second": 3.742,
+      "step": 19300
+    },
+    {
+      "epoch": 8.031627132750728,
+      "eval_international_law_loss": 3.0664596557617188,
+      "eval_international_law_runtime": 0.2666,
+      "eval_international_law_samples_per_second": 7.503,
+      "eval_international_law_steps_per_second": 3.751,
+      "step": 19300
+    },
+    {
+      "epoch": 8.039950062421973,
+      "grad_norm": 0.23046875,
+      "learning_rate": 7.047183738293933e-07,
+      "loss": 0.2291,
+      "step": 19320
+    },
+    {
+      "epoch": 8.048272992093217,
+      "grad_norm": 0.294921875,
+      "learning_rate": 6.990415425493039e-07,
+      "loss": 0.2317,
+      "step": 19340
+    },
+    {
+      "epoch": 8.056595921764462,
+      "grad_norm": 0.28125,
+      "learning_rate": 6.933839499606709e-07,
+      "loss": 0.2325,
+      "step": 19360
+    },
+    {
+      "epoch": 8.064918851435705,
+      "grad_norm": 0.26171875,
+      "learning_rate": 6.877456565005783e-07,
+      "loss": 0.2344,
+      "step": 19380
+    },
+    {
+      "epoch": 8.07324178110695,
+      "grad_norm": 0.244140625,
+      "learning_rate": 6.82126722399948e-07,
+      "loss": 0.2284,
+      "step": 19400
+    },
+    {
+      "epoch": 8.07324178110695,
+      "eval_main_loss": 0.23833835124969482,
+      "eval_main_runtime": 6.3288,
+      "eval_main_samples_per_second": 30.021,
+      "eval_main_steps_per_second": 3.792,
+      "step": 19400
+    },
+    {
+      "epoch": 8.07324178110695,
+      "eval_anatomy_loss": 2.831571102142334,
+      "eval_anatomy_runtime": 0.266,
+      "eval_anatomy_samples_per_second": 7.519,
+      "eval_anatomy_steps_per_second": 3.76,
+      "step": 19400
+    },
+    {
+      "epoch": 8.07324178110695,
+      "eval_college_mathematics_loss": 2.052187442779541,
+      "eval_college_mathematics_runtime": 0.2652,
+      "eval_college_mathematics_samples_per_second": 7.542,
+      "eval_college_mathematics_steps_per_second": 3.771,
+      "step": 19400
+    },
+    {
+      "epoch": 8.07324178110695,
+      "eval_international_law_loss": 3.0645389556884766,
+      "eval_international_law_runtime": 0.2661,
+      "eval_international_law_samples_per_second": 7.516,
+      "eval_international_law_steps_per_second": 3.758,
+      "step": 19400
+    },
+    {
+      "epoch": 8.081564710778194,
+      "grad_norm": 0.25390625,
+      "learning_rate": 6.765272076828961e-07,
+      "loss": 0.2323,
+      "step": 19420
+    },
+    {
+      "epoch": 8.089887640449438,
+      "grad_norm": 0.265625,
+      "learning_rate": 6.709471721660904e-07,
+      "loss": 0.2336,
+      "step": 19440
+    },
+    {
+      "epoch": 8.098210570120683,
+      "grad_norm": 0.294921875,
+      "learning_rate": 6.653866754581159e-07,
+      "loss": 0.2299,
+      "step": 19460
+    },
+    {
+      "epoch": 8.106533499791928,
+      "grad_norm": 0.30859375,
+      "learning_rate": 6.598457769588315e-07,
+      "loss": 0.2282,
+      "step": 19480
+    },
+    {
+      "epoch": 8.11485642946317,
+      "grad_norm": 0.23828125,
+      "learning_rate": 6.5432453585874e-07,
+      "loss": 0.2324,
+      "step": 19500
+    },
+    {
+      "epoch": 8.11485642946317,
+      "eval_main_loss": 0.23833340406417847,
+      "eval_main_runtime": 6.3274,
+      "eval_main_samples_per_second": 30.028,
+      "eval_main_steps_per_second": 3.793,
+      "step": 19500
+    },
+    {
+      "epoch": 8.11485642946317,
+      "eval_anatomy_loss": 2.832037925720215,
+      "eval_anatomy_runtime": 0.2662,
+      "eval_anatomy_samples_per_second": 7.514,
+      "eval_anatomy_steps_per_second": 3.757,
+      "step": 19500
+    },
+    {
+      "epoch": 8.11485642946317,
+      "eval_college_mathematics_loss": 2.0525174140930176,
+      "eval_college_mathematics_runtime": 0.2659,
+      "eval_college_mathematics_samples_per_second": 7.521,
+      "eval_college_mathematics_steps_per_second": 3.76,
+      "step": 19500
+    },
+    {
+      "epoch": 8.11485642946317,
+      "eval_international_law_loss": 3.0670900344848633,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.506,
+      "eval_international_law_steps_per_second": 3.753,
+      "step": 19500
+    },
+    {
+      "epoch": 8.123179359134415,
+      "grad_norm": 0.25390625,
+      "learning_rate": 6.488230111383553e-07,
+      "loss": 0.2317,
+      "step": 19520
+    },
+    {
+      "epoch": 8.13150228880566,
+      "grad_norm": 0.3125,
+      "learning_rate": 6.433412615675705e-07,
+      "loss": 0.2324,
+      "step": 19540
+    },
+    {
+      "epoch": 8.139825218476904,
+      "grad_norm": 0.265625,
+      "learning_rate": 6.378793457050306e-07,
+      "loss": 0.234,
+      "step": 19560
+    },
+    {
+      "epoch": 8.148148148148149,
+      "grad_norm": 0.2451171875,
+      "learning_rate": 6.324373218975105e-07,
+      "loss": 0.234,
+      "step": 19580
+    },
+    {
+      "epoch": 8.156471077819392,
+      "grad_norm": 0.283203125,
+      "learning_rate": 6.270152482792844e-07,
+      "loss": 0.2307,
+      "step": 19600
+    },
+    {
+      "epoch": 8.156471077819392,
+      "eval_main_loss": 0.23835553228855133,
+      "eval_main_runtime": 6.3341,
+      "eval_main_samples_per_second": 29.996,
+      "eval_main_steps_per_second": 3.789,
+      "step": 19600
+    },
+    {
+      "epoch": 8.156471077819392,
+      "eval_anatomy_loss": 2.833484649658203,
+      "eval_anatomy_runtime": 0.2668,
+      "eval_anatomy_samples_per_second": 7.496,
+      "eval_anatomy_steps_per_second": 3.748,
+      "step": 19600
+    },
+    {
+      "epoch": 8.156471077819392,
+      "eval_college_mathematics_loss": 2.054138422012329,
+      "eval_college_mathematics_runtime": 0.2646,
+      "eval_college_mathematics_samples_per_second": 7.558,
+      "eval_college_mathematics_steps_per_second": 3.779,
+      "step": 19600
+    },
+    {
+      "epoch": 8.156471077819392,
+      "eval_international_law_loss": 3.066187620162964,
+      "eval_international_law_runtime": 0.2674,
+      "eval_international_law_samples_per_second": 7.478,
+      "eval_international_law_steps_per_second": 3.739,
+      "step": 19600
+    },
+    {
+      "epoch": 8.164794007490636,
+      "grad_norm": 0.30859375,
+      "learning_rate": 6.21613182771513e-07,
+      "loss": 0.2277,
+      "step": 19620
+    },
+    {
+      "epoch": 8.17311693716188,
+      "grad_norm": 0.26953125,
+      "learning_rate": 6.162311830816187e-07,
+      "loss": 0.2276,
+      "step": 19640
+    },
+    {
+      "epoch": 8.181439866833125,
+      "grad_norm": 0.31640625,
+      "learning_rate": 6.108693067026713e-07,
+      "loss": 0.2329,
+      "step": 19660
+    },
+    {
+      "epoch": 8.18976279650437,
+      "grad_norm": 0.291015625,
+      "learning_rate": 6.05527610912777e-07,
+      "loss": 0.2341,
+      "step": 19680
+    },
+    {
+      "epoch": 8.198085726175615,
+      "grad_norm": 0.28515625,
+      "learning_rate": 6.002061527744573e-07,
+      "loss": 0.2313,
+      "step": 19700
+    },
+    {
+      "epoch": 8.198085726175615,
+      "eval_main_loss": 0.23825512826442719,
+      "eval_main_runtime": 6.325,
+      "eval_main_samples_per_second": 30.04,
+      "eval_main_steps_per_second": 3.794,
+      "step": 19700
+    },
+    {
+      "epoch": 8.198085726175615,
+      "eval_anatomy_loss": 2.831984519958496,
+      "eval_anatomy_runtime": 0.2658,
+      "eval_anatomy_samples_per_second": 7.523,
+      "eval_anatomy_steps_per_second": 3.762,
+      "step": 19700
+    },
+    {
+      "epoch": 8.198085726175615,
+      "eval_college_mathematics_loss": 2.0513973236083984,
+      "eval_college_mathematics_runtime": 0.2657,
+      "eval_college_mathematics_samples_per_second": 7.528,
+      "eval_college_mathematics_steps_per_second": 3.764,
+      "step": 19700
+    },
+    {
+      "epoch": 8.198085726175615,
+      "eval_international_law_loss": 3.0651326179504395,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.506,
+      "eval_international_law_steps_per_second": 3.753,
+      "step": 19700
+    },
+    {
+      "epoch": 8.206408655846857,
+      "grad_norm": 0.32421875,
+      "learning_rate": 5.94904989134052e-07,
+      "loss": 0.2367,
+      "step": 19720
+    },
+    {
+      "epoch": 8.214731585518102,
+      "grad_norm": 0.296875,
+      "learning_rate": 5.896241766211011e-07,
+      "loss": 0.232,
+      "step": 19740
+    },
+    {
+      "epoch": 8.223054515189347,
+      "grad_norm": 0.31640625,
+      "learning_rate": 5.843637716477454e-07,
+      "loss": 0.2336,
+      "step": 19760
+    },
+    {
+      "epoch": 8.231377444860591,
+      "grad_norm": 0.318359375,
+      "learning_rate": 5.791238304081245e-07,
+      "loss": 0.235,
+      "step": 19780
+    },
+    {
+      "epoch": 8.239700374531836,
+      "grad_norm": 0.26953125,
+      "learning_rate": 5.73904408877772e-07,
+      "loss": 0.2298,
+      "step": 19800
+    },
+    {
+      "epoch": 8.239700374531836,
+      "eval_main_loss": 0.23827609419822693,
+      "eval_main_runtime": 6.324,
+      "eval_main_samples_per_second": 30.044,
+      "eval_main_steps_per_second": 3.795,
+      "step": 19800
+    },
+    {
+      "epoch": 8.239700374531836,
+      "eval_anatomy_loss": 2.8319296836853027,
+      "eval_anatomy_runtime": 0.2663,
+      "eval_anatomy_samples_per_second": 7.51,
+      "eval_anatomy_steps_per_second": 3.755,
+      "step": 19800
+    },
+    {
+      "epoch": 8.239700374531836,
+      "eval_college_mathematics_loss": 2.051198720932007,
+      "eval_college_mathematics_runtime": 0.2656,
+      "eval_college_mathematics_samples_per_second": 7.531,
+      "eval_college_mathematics_steps_per_second": 3.765,
+      "step": 19800
+    },
+    {
+      "epoch": 8.239700374531836,
+      "eval_international_law_loss": 3.0645523071289062,
+      "eval_international_law_runtime": 0.2649,
+      "eval_international_law_samples_per_second": 7.551,
+      "eval_international_law_steps_per_second": 3.775,
+      "step": 19800
+    },
+    {
+      "epoch": 8.24802330420308,
+      "grad_norm": 0.29296875,
+      "learning_rate": 5.687055628130219e-07,
+      "loss": 0.2318,
+      "step": 19820
+    },
+    {
+      "epoch": 8.256346233874323,
+      "grad_norm": 0.2373046875,
+      "learning_rate": 5.6352734775041e-07,
+      "loss": 0.231,
+      "step": 19840
+    },
+    {
+      "epoch": 8.264669163545568,
+      "grad_norm": 0.27734375,
+      "learning_rate": 5.58369819006084e-07,
+      "loss": 0.2331,
+      "step": 19860
+    },
+    {
+      "epoch": 8.272992093216812,
+      "grad_norm": 0.33203125,
+      "learning_rate": 5.532330316752091e-07,
+      "loss": 0.2343,
+      "step": 19880
+    },
+    {
+      "epoch": 8.281315022888057,
+      "grad_norm": 0.298828125,
+      "learning_rate": 5.481170406313799e-07,
+      "loss": 0.2327,
+      "step": 19900
+    },
+    {
+      "epoch": 8.281315022888057,
+      "eval_main_loss": 0.23821409046649933,
+      "eval_main_runtime": 6.3564,
+      "eval_main_samples_per_second": 29.891,
+      "eval_main_steps_per_second": 3.776,
+      "step": 19900
+    },
+    {
+      "epoch": 8.281315022888057,
+      "eval_anatomy_loss": 2.8317759037017822,
+      "eval_anatomy_runtime": 0.2667,
+      "eval_anatomy_samples_per_second": 7.5,
+      "eval_anatomy_steps_per_second": 3.75,
+      "step": 19900
+    },
+    {
+      "epoch": 8.281315022888057,
+      "eval_college_mathematics_loss": 2.051119804382324,
+      "eval_college_mathematics_runtime": 0.2668,
+      "eval_college_mathematics_samples_per_second": 7.497,
+      "eval_college_mathematics_steps_per_second": 3.748,
+      "step": 19900
+    },
+    {
+      "epoch": 8.281315022888057,
+      "eval_international_law_loss": 3.0645089149475098,
+      "eval_international_law_runtime": 0.2665,
+      "eval_international_law_samples_per_second": 7.506,
+      "eval_international_law_steps_per_second": 3.753,
+      "step": 19900
+    },
+    {
+      "epoch": 8.289637952559302,
+      "grad_norm": 0.28125,
+      "learning_rate": 5.430219005260387e-07,
+      "loss": 0.2335,
+      "step": 19920
+    },
+    {
+      "epoch": 8.297960882230544,
+      "grad_norm": 0.201171875,
+      "learning_rate": 5.379476657878834e-07,
+      "loss": 0.2349,
+      "step": 19940
+    },
+    {
+      "epoch": 8.306283811901789,
+      "grad_norm": 0.302734375,
+      "learning_rate": 5.328943906222955e-07,
+      "loss": 0.2327,
+      "step": 19960
+    },
+    {
+      "epoch": 8.314606741573034,
+      "grad_norm": 0.25390625,
+      "learning_rate": 5.278621290107533e-07,
+      "loss": 0.2318,
+      "step": 19980
+    },
+    {
+      "epoch": 8.322929671244278,
+      "grad_norm": 0.26171875,
+      "learning_rate": 5.228509347102593e-07,
+      "loss": 0.2314,
+      "step": 20000
+    },
+    {
+      "epoch": 8.322929671244278,
+      "eval_main_loss": 0.2383708357810974,
+      "eval_main_runtime": 6.3506,
+      "eval_main_samples_per_second": 29.918,
+      "eval_main_steps_per_second": 3.779,
+      "step": 20000
+    },
+    {
+      "epoch": 8.322929671244278,
+      "eval_anatomy_loss": 2.8334150314331055,
+      "eval_anatomy_runtime": 0.2681,
+      "eval_anatomy_samples_per_second": 7.459,
+      "eval_anatomy_steps_per_second": 3.73,
+      "step": 20000
+    },
+    {
+      "epoch": 8.322929671244278,
+      "eval_college_mathematics_loss": 2.049489974975586,
+      "eval_college_mathematics_runtime": 0.2668,
+      "eval_college_mathematics_samples_per_second": 7.496,
+      "eval_college_mathematics_steps_per_second": 3.748,
+      "step": 20000
+    },
+    {
+      "epoch": 8.322929671244278,
+      "eval_international_law_loss": 3.0655791759490967,
+      "eval_international_law_runtime": 0.2674,
+      "eval_international_law_samples_per_second": 7.48,
+      "eval_international_law_steps_per_second": 3.74,
+      "step": 20000
+    },
+    {
+      "epoch": 8.331252600915523,
+      "grad_norm": 0.30859375,
+      "learning_rate": 5.178608612527663e-07,
+      "loss": 0.2332,
+      "step": 20020
+    },
+    {
+      "epoch": 8.339575530586767,
+      "grad_norm": 0.2431640625,
+      "learning_rate": 5.12891961944601e-07,
+      "loss": 0.2299,
+      "step": 20040
+    },
+    {
+      "epoch": 8.34789846025801,
+      "grad_norm": 0.279296875,
+      "learning_rate": 5.079442898659017e-07,
+      "loss": 0.2326,
+      "step": 20060
+    },
+    {
+      "epoch": 8.356221389929255,
+      "grad_norm": 0.265625,
+      "learning_rate": 5.030178978700448e-07,
+      "loss": 0.2295,
+      "step": 20080
+    },
+    {
+      "epoch": 8.3645443196005,
+      "grad_norm": 0.24609375,
+      "learning_rate": 4.98112838583083e-07,
+      "loss": 0.2306,
+      "step": 20100
+    },
+    {
+      "epoch": 8.3645443196005,
+      "eval_main_loss": 0.23833905160427094,
+      "eval_main_runtime": 6.343,
+      "eval_main_samples_per_second": 29.954,
+      "eval_main_steps_per_second": 3.784,
+      "step": 20100
+    },
+    {
+      "epoch": 8.3645443196005,
+      "eval_anatomy_loss": 2.832882881164551,
+      "eval_anatomy_runtime": 0.2664,
+      "eval_anatomy_samples_per_second": 7.508,
+      "eval_anatomy_steps_per_second": 3.754,
+      "step": 20100
+    },
+    {
+      "epoch": 8.3645443196005,
+      "eval_college_mathematics_loss": 2.0530946254730225,
+      "eval_college_mathematics_runtime": 0.2677,
+      "eval_college_mathematics_samples_per_second": 7.472,
+      "eval_college_mathematics_steps_per_second": 3.736,
+      "step": 20100
+    },
+    {
+      "epoch": 8.3645443196005,
+      "eval_international_law_loss": 3.066558599472046,
+      "eval_international_law_runtime": 0.2659,
+      "eval_international_law_samples_per_second": 7.522,
+      "eval_international_law_steps_per_second": 3.761,
+      "step": 20100
+    },
+    {
+      "epoch": 8.372867249271744,
+      "grad_norm": 0.29296875,
+      "learning_rate": 4.932291644031844e-07,
+      "loss": 0.232,
+      "step": 20120
+    },
+    {
+      "epoch": 8.381190178942989,
+      "grad_norm": 0.287109375,
+      "learning_rate": 4.883669275000699e-07,
+      "loss": 0.2289,
+      "step": 20140
+    },
+    {
+      "epoch": 8.389513108614231,
+      "grad_norm": 0.265625,
+      "learning_rate": 4.835261798144569e-07,
+      "loss": 0.2359,
+      "step": 20160
+    },
+    {
+      "epoch": 8.397836038285476,
+      "grad_norm": 0.25390625,
+      "learning_rate": 4.787069730575067e-07,
+      "loss": 0.2288,
+      "step": 20180
+    },
+    {
+      "epoch": 8.40615896795672,
+      "grad_norm": 0.267578125,
+      "learning_rate": 4.739093587102686e-07,
+      "loss": 0.2314,
+      "step": 20200
+    },
+    {
+      "epoch": 8.40615896795672,
+      "eval_main_loss": 0.23831138014793396,
+      "eval_main_runtime": 6.3296,
+      "eval_main_samples_per_second": 30.018,
+      "eval_main_steps_per_second": 3.792,
+      "step": 20200
+    },
+    {
+      "epoch": 8.40615896795672,
+      "eval_anatomy_loss": 2.833329916000366,
+      "eval_anatomy_runtime": 0.266,
+      "eval_anatomy_samples_per_second": 7.519,
+      "eval_anatomy_steps_per_second": 3.759,
+      "step": 20200
+    },
+    {
+      "epoch": 8.40615896795672,
+      "eval_college_mathematics_loss": 2.0539867877960205,
+      "eval_college_mathematics_runtime": 0.2671,
+      "eval_college_mathematics_samples_per_second": 7.488,
+      "eval_college_mathematics_steps_per_second": 3.744,
+      "step": 20200
+    },
+    {
+      "epoch": 8.40615896795672,
+      "eval_international_law_loss": 3.067291498184204,
+      "eval_international_law_runtime": 0.2661,
+      "eval_international_law_samples_per_second": 7.515,
+      "eval_international_law_steps_per_second": 3.757,
+      "step": 20200
+    },
+    {
+      "epoch": 8.414481897627965,
+      "grad_norm": 0.232421875,
+      "learning_rate": 4.6913338802313165e-07,
+      "loss": 0.227,
+      "step": 20220
+    },
+    {
+      "epoch": 8.42280482729921,
+      "grad_norm": 0.318359375,
+      "learning_rate": 4.6437911201527686e-07,
+      "loss": 0.2312,
+      "step": 20240
+    },
+    {
+      "epoch": 8.431127756970454,
+      "grad_norm": 0.330078125,
+      "learning_rate": 4.596465814741341e-07,
+      "loss": 0.2295,
+      "step": 20260
+    },
+    {
+      "epoch": 8.439450686641697,
+      "grad_norm": 0.236328125,
+      "learning_rate": 4.5493584695483593e-07,
+      "loss": 0.2321,
+      "step": 20280
+    },
+    {
+      "epoch": 8.447773616312942,
+      "grad_norm": 0.306640625,
+      "learning_rate": 4.502469587796807e-07,
+      "loss": 0.2317,
+      "step": 20300
+    },
+    {
+      "epoch": 8.447773616312942,
+      "eval_main_loss": 0.23827072978019714,
+      "eval_main_runtime": 6.3259,
+      "eval_main_samples_per_second": 30.035,
+      "eval_main_steps_per_second": 3.794,
+      "step": 20300
+    },
+    {
+      "epoch": 8.447773616312942,
+      "eval_anatomy_loss": 2.834202766418457,
+      "eval_anatomy_runtime": 0.2657,
+      "eval_anatomy_samples_per_second": 7.527,
+      "eval_anatomy_steps_per_second": 3.764,
+      "step": 20300
+    },
+    {
+      "epoch": 8.447773616312942,
+      "eval_college_mathematics_loss": 2.0520122051239014,
+      "eval_college_mathematics_runtime": 0.2664,
+      "eval_college_mathematics_samples_per_second": 7.506,
+      "eval_college_mathematics_steps_per_second": 3.753,
+      "step": 20300
+    },
+    {
+      "epoch": 8.447773616312942,
+      "eval_international_law_loss": 3.065819025039673,
+      "eval_international_law_runtime": 0.266,
+      "eval_international_law_samples_per_second": 7.517,
+      "eval_international_law_steps_per_second": 3.759,
+      "step": 20300
+    },
+    {
+      "epoch": 8.456096545984186,
+      "grad_norm": 0.26171875,
+      "learning_rate": 4.4557996703759295e-07,
+      "loss": 0.2337,
+      "step": 20320
+    },
+    {
+      "epoch": 8.464419475655431,
+      "grad_norm": 0.267578125,
+      "learning_rate": 4.409349215835887e-07,
+      "loss": 0.2304,
+      "step": 20340
+    },
+    {
+      "epoch": 8.472742405326676,
+      "grad_norm": 0.27734375,
+      "learning_rate": 4.363118720382456e-07,
+      "loss": 0.2306,
+      "step": 20360
+    },
+    {
+      "epoch": 8.481065334997918,
+      "grad_norm": 0.263671875,
+      "learning_rate": 4.317108677871687e-07,
+      "loss": 0.2262,
+      "step": 20380
+    },
+    {
+      "epoch": 8.489388264669163,
+      "grad_norm": 0.275390625,
+      "learning_rate": 4.271319579804639e-07,
+      "loss": 0.2297,
+      "step": 20400
+    },
+    {
+      "epoch": 8.489388264669163,
+      "eval_main_loss": 0.23833975195884705,
+      "eval_main_runtime": 6.3573,
+      "eval_main_samples_per_second": 29.887,
+      "eval_main_steps_per_second": 3.775,
+      "step": 20400
+    },
+    {
+      "epoch": 8.489388264669163,
+      "eval_anatomy_loss": 2.8326468467712402,
+      "eval_anatomy_runtime": 0.2676,
+      "eval_anatomy_samples_per_second": 7.475,
+      "eval_anatomy_steps_per_second": 3.738,
+      "step": 20400
+    },
+    {
+      "epoch": 8.489388264669163,
+      "eval_college_mathematics_loss": 2.0523431301116943,
+      "eval_college_mathematics_runtime": 0.2667,
+      "eval_college_mathematics_samples_per_second": 7.5,
+      "eval_college_mathematics_steps_per_second": 3.75,
+      "step": 20400
+    },
+    {
+      "epoch": 8.489388264669163,
+      "eval_international_law_loss": 3.0675792694091797,
+      "eval_international_law_runtime": 0.2662,
+      "eval_international_law_samples_per_second": 7.512,
+      "eval_international_law_steps_per_second": 3.756,
+      "step": 20400
+    },
+    {
+      "epoch": 8.497711194340408,
+      "grad_norm": 0.26171875,
+      "learning_rate": 4.2257519153221736e-07,
+      "loss": 0.2366,
+      "step": 20420
+    },
+    {
+      "epoch": 8.506034124011652,
+      "grad_norm": 0.349609375,
+      "learning_rate": 4.180406171199644e-07,
+      "loss": 0.2301,
+      "step": 20440
+    },
+    {
+      "epoch": 8.514357053682897,
+      "grad_norm": 0.2734375,
+      "learning_rate": 4.1352828318417915e-07,
+      "loss": 0.2337,
+      "step": 20460
+    },
+    {
+      "epoch": 8.522679983354141,
+      "grad_norm": 0.265625,
+      "learning_rate": 4.090382379277499e-07,
+      "loss": 0.2325,
+      "step": 20480
+    },
+    {
+      "epoch": 8.531002913025384,
+      "grad_norm": 0.255859375,
+      "learning_rate": 4.045705293154664e-07,
+      "loss": 0.2335,
+      "step": 20500
+    },
+    {
+      "epoch": 8.531002913025384,
+      "eval_main_loss": 0.2383665144443512,
+      "eval_main_runtime": 6.3513,
+      "eval_main_samples_per_second": 29.915,
+      "eval_main_steps_per_second": 3.779,
+      "step": 20500
+    },
+    {
+      "epoch": 8.531002913025384,
+      "eval_anatomy_loss": 2.8339719772338867,
+      "eval_anatomy_runtime": 0.2664,
+      "eval_anatomy_samples_per_second": 7.507,
+      "eval_anatomy_steps_per_second": 3.753,
+      "step": 20500
+    },
+    {
+      "epoch": 8.531002913025384,
+      "eval_college_mathematics_loss": 2.0505967140197754,
+      "eval_college_mathematics_runtime": 0.2672,
+      "eval_college_mathematics_samples_per_second": 7.486,
+      "eval_college_mathematics_steps_per_second": 3.743,
+      "step": 20500
+    },
+    {
+      "epoch": 8.531002913025384,
+      "eval_international_law_loss": 3.065483808517456,
+      "eval_international_law_runtime": 0.2682,
+      "eval_international_law_samples_per_second": 7.457,
+      "eval_international_law_steps_per_second": 3.729,
+      "step": 20500
+    },
+    {
+      "epoch": 8.539325842696629,
+      "grad_norm": 0.251953125,
+      "learning_rate": 4.001252050735102e-07,
+      "loss": 0.2313,
+      "step": 20520
+    },
+    {
+      "epoch": 8.547648772367873,
+      "grad_norm": 0.2109375,
+      "learning_rate": 3.9570231268893975e-07,
+      "loss": 0.2308,
+      "step": 20540
+    },
+    {
+      "epoch": 8.555971702039118,
+      "grad_norm": 0.314453125,
+      "learning_rate": 3.9130189940918745e-07,
+      "loss": 0.2313,
+      "step": 20560
+    },
+    {
+      "epoch": 8.564294631710363,
+      "grad_norm": 0.310546875,
+      "learning_rate": 3.869240122415521e-07,
+      "loss": 0.233,
+      "step": 20580
+    },
+    {
+      "epoch": 8.572617561381607,
+      "grad_norm": 0.2265625,
+      "learning_rate": 3.8256869795269824e-07,
+      "loss": 0.2285,
+      "step": 20600
+    },
+    {
+      "epoch": 8.572617561381607,
+      "eval_main_loss": 0.2381974309682846,
+      "eval_main_runtime": 6.3331,
+      "eval_main_samples_per_second": 30.001,
+      "eval_main_steps_per_second": 3.79,
+      "step": 20600
+    },
+    {
+      "epoch": 8.572617561381607,
+      "eval_anatomy_loss": 2.835426092147827,
+      "eval_anatomy_runtime": 0.2654,
+      "eval_anatomy_samples_per_second": 7.537,
+      "eval_anatomy_steps_per_second": 3.769,
+      "step": 20600
+    },
+    {
+      "epoch": 8.572617561381607,
+      "eval_college_mathematics_loss": 2.0525310039520264,
+      "eval_college_mathematics_runtime": 0.2646,
+      "eval_college_mathematics_samples_per_second": 7.559,
+      "eval_college_mathematics_steps_per_second": 3.78,
+      "step": 20600
+    },
+    {
+      "epoch": 8.572617561381607,
+      "eval_international_law_loss": 3.064969539642334,
+      "eval_international_law_runtime": 0.265,
+      "eval_international_law_samples_per_second": 7.547,
+      "eval_international_law_steps_per_second": 3.774,
+      "step": 20600
+    },
+    {
+      "epoch": 8.58094049105285,
+      "grad_norm": 0.2578125,
+      "learning_rate": 3.782360030681578e-07,
+      "loss": 0.2336,
+      "step": 20620
+    },
+    {
+      "epoch": 8.589263420724095,
+      "grad_norm": 0.283203125,
+      "learning_rate": 3.73925973871829e-07,
+      "loss": 0.2336,
+      "step": 20640
+    },
+    {
+      "epoch": 8.59758635039534,
+      "grad_norm": 0.314453125,
+      "learning_rate": 3.696386564054863e-07,
+      "loss": 0.2337,
+      "step": 20660
+    },
+    {
+      "epoch": 8.605909280066584,
+      "grad_norm": 0.2412109375,
+      "learning_rate": 3.6537409646828505e-07,
+      "loss": 0.2329,
+      "step": 20680
+    },
+    {
+      "epoch": 8.614232209737828,
+      "grad_norm": 0.248046875,
+      "learning_rate": 3.611323396162758e-07,
+      "loss": 0.2354,
+      "step": 20700
+    },
+    {
+      "epoch": 8.614232209737828,
+      "eval_main_loss": 0.23833750188350677,
+      "eval_main_runtime": 6.3553,
+      "eval_main_samples_per_second": 29.896,
+      "eval_main_steps_per_second": 3.776,
+      "step": 20700
+    },
+    {
+      "epoch": 8.614232209737828,
+      "eval_anatomy_loss": 2.8340907096862793,
+      "eval_anatomy_runtime": 0.2683,
+      "eval_anatomy_samples_per_second": 7.455,
+      "eval_anatomy_steps_per_second": 3.728,
+      "step": 20700
+    },
+    {
+      "epoch": 8.614232209737828,
+      "eval_college_mathematics_loss": 2.0527515411376953,
+      "eval_college_mathematics_runtime": 0.2665,
+      "eval_college_mathematics_samples_per_second": 7.504,
+      "eval_college_mathematics_steps_per_second": 3.752,
+      "step": 20700
+    },
+    {
+      "epoch": 8.614232209737828,
+      "eval_international_law_loss": 3.065247058868408,
+      "eval_international_law_runtime": 0.2686,
+      "eval_international_law_samples_per_second": 7.447,
+      "eval_international_law_steps_per_second": 3.724,
+      "step": 20700
+    },
+    {
+      "epoch": 8.622555139409073,
+      "grad_norm": 0.27734375,
+      "learning_rate": 3.569134311619146e-07,
+      "loss": 0.2259,
+      "step": 20720
+    },
+    {
+      "epoch": 8.630878069080316,
+      "grad_norm": 0.26953125,
+      "learning_rate": 3.527174161735797e-07,
+      "loss": 0.2324,
+      "step": 20740
+    },
+    {
+      "epoch": 8.63920099875156,
+      "grad_norm": 0.28125,
+      "learning_rate": 3.4854433947509256e-07,
+      "loss": 0.2314,
+      "step": 20760
+    },
+    {
+      "epoch": 8.647523928422805,
+      "grad_norm": 0.28125,
+      "learning_rate": 3.4439424564523346e-07,
+      "loss": 0.2357,
+      "step": 20780
+    },
+    {
+      "epoch": 8.65584685809405,
+      "grad_norm": 0.234375,
+      "learning_rate": 3.402671790172718e-07,
+      "loss": 0.2318,
+      "step": 20800
+    },
+    {
+      "epoch": 8.65584685809405,
+      "eval_main_loss": 0.23831304907798767,
+      "eval_main_runtime": 6.3537,
+      "eval_main_samples_per_second": 29.904,
+      "eval_main_steps_per_second": 3.777,
+      "step": 20800
+    },
+    {
+      "epoch": 8.65584685809405,
+      "eval_anatomy_loss": 2.8343677520751953,
+      "eval_anatomy_runtime": 0.2677,
+      "eval_anatomy_samples_per_second": 7.472,
+      "eval_anatomy_steps_per_second": 3.736,
+      "step": 20800
+    },
+    {
+      "epoch": 8.65584685809405,
+      "eval_college_mathematics_loss": 2.0545074939727783,
+      "eval_college_mathematics_runtime": 0.2666,
+      "eval_college_mathematics_samples_per_second": 7.501,
+      "eval_college_mathematics_steps_per_second": 3.751,
+      "step": 20800
+    },
+    {
+      "epoch": 8.65584685809405,
+      "eval_international_law_loss": 3.0653839111328125,
+      "eval_international_law_runtime": 0.2681,
+      "eval_international_law_samples_per_second": 7.461,
+      "eval_international_law_steps_per_second": 3.731,
+      "step": 20800
+    },
+    {
+      "epoch": 8.664169787765294,
+      "grad_norm": 0.333984375,
+      "learning_rate": 3.361631836784898e-07,
+      "loss": 0.2319,
+      "step": 20820
+    },
+    {
+      "epoch": 8.672492717436537,
+      "grad_norm": 0.291015625,
+      "learning_rate": 3.320823034697074e-07,
+      "loss": 0.2332,
+      "step": 20840
+    },
+    {
+      "epoch": 8.680815647107782,
+      "grad_norm": 0.28515625,
+      "learning_rate": 3.280245819848224e-07,
+      "loss": 0.2346,
+      "step": 20860
+    },
+    {
+      "epoch": 8.689138576779026,
+      "grad_norm": 0.2333984375,
+      "learning_rate": 3.239900625703374e-07,
+      "loss": 0.2327,
+      "step": 20880
+    },
+    {
+      "epoch": 8.69746150645027,
+      "grad_norm": 0.291015625,
+      "learning_rate": 3.199787883248992e-07,
+      "loss": 0.2333,
+      "step": 20900
+    },
+    {
+      "epoch": 8.69746150645027,
+      "eval_main_loss": 0.23826073110103607,
+      "eval_main_runtime": 6.3163,
+      "eval_main_samples_per_second": 30.081,
+      "eval_main_steps_per_second": 3.8,
+      "step": 20900
+    },
+    {
+      "epoch": 8.69746150645027,
+      "eval_anatomy_loss": 2.8350343704223633,
+      "eval_anatomy_runtime": 0.2662,
+      "eval_anatomy_samples_per_second": 7.514,
+      "eval_anatomy_steps_per_second": 3.757,
+      "step": 20900
+    },
+    {
+      "epoch": 8.69746150645027,
+      "eval_college_mathematics_loss": 2.052778959274292,
+      "eval_college_mathematics_runtime": 0.2653,
+      "eval_college_mathematics_samples_per_second": 7.538,
+      "eval_college_mathematics_steps_per_second": 3.769,
+      "step": 20900
+    },
+    {
+      "epoch": 8.69746150645027,
+      "eval_international_law_loss": 3.064143180847168,
+      "eval_international_law_runtime": 0.2658,
+      "eval_international_law_samples_per_second": 7.523,
+      "eval_international_law_steps_per_second": 3.762,
+      "step": 20900
+    },
+    {
+      "epoch": 8.705784436121515,
+      "grad_norm": 0.24609375,
+      "learning_rate": 3.15990802098842e-07,
+      "loss": 0.2307,
+      "step": 20920
+    },
+    {
+      "epoch": 8.71410736579276,
+      "grad_norm": 0.2431640625,
+      "learning_rate": 3.12026146493721e-07,
+      "loss": 0.2292,
+      "step": 20940
+    },
+    {
+      "epoch": 8.722430295464003,
+      "grad_norm": 0.30078125,
+      "learning_rate": 3.0808486386186804e-07,
+      "loss": 0.2302,
+      "step": 20960
+    },
+    {
+      "epoch": 8.730753225135247,
+      "grad_norm": 0.271484375,
+      "learning_rate": 3.041669963059304e-07,
+      "loss": 0.2284,
+      "step": 20980
+    },
+    {
+      "epoch": 8.739076154806492,
+      "grad_norm": 0.31640625,
+      "learning_rate": 3.0027258567842525e-07,
+      "loss": 0.2315,
+      "step": 21000
+    },
+    {
+      "epoch": 8.739076154806492,
+      "eval_main_loss": 0.23824970424175262,
+      "eval_main_runtime": 6.3549,
+      "eval_main_samples_per_second": 29.898,
+      "eval_main_steps_per_second": 3.777,
+      "step": 21000
+    },
+    {
+      "epoch": 8.739076154806492,
+      "eval_anatomy_loss": 2.8310306072235107,
+      "eval_anatomy_runtime": 0.2668,
+      "eval_anatomy_samples_per_second": 7.497,
+      "eval_anatomy_steps_per_second": 3.749,
+      "step": 21000
+    },
+    {
+      "epoch": 8.739076154806492,
+      "eval_college_mathematics_loss": 2.0544137954711914,
+      "eval_college_mathematics_runtime": 0.2671,
+      "eval_college_mathematics_samples_per_second": 7.488,
+      "eval_college_mathematics_steps_per_second": 3.744,
+      "step": 21000
+    },
+    {
+      "epoch": 8.739076154806492,
+      "eval_international_law_loss": 3.0671803951263428,
+      "eval_international_law_runtime": 0.2679,
+      "eval_international_law_samples_per_second": 7.465,
+      "eval_international_law_steps_per_second": 3.733,
+      "step": 21000
+    },
+    {
+      "epoch": 8.747399084477737,
+      "grad_norm": 0.28515625,
+      "learning_rate": 2.9640167358129273e-07,
+      "loss": 0.2344,
+      "step": 21020
+    },
+    {
+      "epoch": 8.755722014148981,
+      "grad_norm": 0.30859375,
+      "learning_rate": 2.9255430136544886e-07,
+      "loss": 0.231,
+      "step": 21040
+    },
+    {
+      "epoch": 8.764044943820224,
+      "grad_norm": 0.3203125,
+      "learning_rate": 2.8873051013034695e-07,
+      "loss": 0.2341,
+      "step": 21060
+    },
+    {
+      "epoch": 8.772367873491469,
+      "grad_norm": 0.25,
+      "learning_rate": 2.84930340723536e-07,
+      "loss": 0.229,
+      "step": 21080
+    },
+    {
+      "epoch": 8.780690803162713,
+      "grad_norm": 0.275390625,
+      "learning_rate": 2.811538337402264e-07,
+      "loss": 0.2285,
+      "step": 21100
+    },
+    {
+      "epoch": 8.780690803162713,
+      "eval_main_loss": 0.23829315602779388,
+      "eval_main_runtime": 6.3571,
+      "eval_main_samples_per_second": 29.888,
+      "eval_main_steps_per_second": 3.775,
+      "step": 21100
+    },
+    {
+      "epoch": 8.780690803162713,
+      "eval_anatomy_loss": 2.8332481384277344,
+      "eval_anatomy_runtime": 0.2673,
+      "eval_anatomy_samples_per_second": 7.483,
+      "eval_anatomy_steps_per_second": 3.742,
+      "step": 21100
+    },
+    {
+      "epoch": 8.780690803162713,
+      "eval_college_mathematics_loss": 2.0525107383728027,
+      "eval_college_mathematics_runtime": 0.2662,
+      "eval_college_mathematics_samples_per_second": 7.512,
+      "eval_college_mathematics_steps_per_second": 3.756,
+      "step": 21100
+    },
+    {
+      "epoch": 8.780690803162713,
+      "eval_international_law_loss": 3.0667715072631836,
+      "eval_international_law_runtime": 0.2664,
+      "eval_international_law_samples_per_second": 7.507,
+      "eval_international_law_steps_per_second": 3.754,
+      "step": 21100
+    },
+    {
+      "epoch": 8.789013732833958,
+      "grad_norm": 0.306640625,
+      "learning_rate": 2.774010295228549e-07,
+      "loss": 0.2297,
+      "step": 21120
+    },
+    {
+      "epoch": 8.797336662505202,
+      "grad_norm": 0.283203125,
+      "learning_rate": 2.7367196816065397e-07,
+      "loss": 0.2364,
+      "step": 21140
+    },
+    {
+      "epoch": 8.805659592176447,
+      "grad_norm": 0.283203125,
+      "learning_rate": 2.699666894892236e-07,
+      "loss": 0.2285,
+      "step": 21160
+    },
+    {
+      "epoch": 8.81398252184769,
+      "grad_norm": 0.291015625,
+      "learning_rate": 2.662852330901053e-07,
+      "loss": 0.24,
+      "step": 21180
+    },
+    {
+      "epoch": 8.822305451518934,
+      "grad_norm": 0.27734375,
+      "learning_rate": 2.6262763829036197e-07,
+      "loss": 0.2316,
+      "step": 21200
+    },
+    {
+      "epoch": 8.822305451518934,
+      "eval_main_loss": 0.23829114437103271,
+      "eval_main_runtime": 6.3393,
+      "eval_main_samples_per_second": 29.972,
+      "eval_main_steps_per_second": 3.786,
+      "step": 21200
+    },
+    {
+      "epoch": 8.822305451518934,
+      "eval_anatomy_loss": 2.832653045654297,
+      "eval_anatomy_runtime": 0.2661,
+      "eval_anatomy_samples_per_second": 7.515,
+      "eval_anatomy_steps_per_second": 3.758,
+      "step": 21200
+    },
+    {
+      "epoch": 8.822305451518934,
+      "eval_college_mathematics_loss": 2.0520215034484863,
+      "eval_college_mathematics_runtime": 0.2666,
+      "eval_college_mathematics_samples_per_second": 7.502,
+      "eval_college_mathematics_steps_per_second": 3.751,
+      "step": 21200
+    },
+    {
+      "epoch": 8.822305451518934,
+      "eval_international_law_loss": 3.0661661624908447,
+      "eval_international_law_runtime": 0.2661,
+      "eval_international_law_samples_per_second": 7.515,
+      "eval_international_law_steps_per_second": 3.757,
+      "step": 21200
+    },
+    {
+      "epoch": 8.830628381190179,
+      "grad_norm": 0.2734375,
+      "learning_rate": 2.5899394416215305e-07,
+      "loss": 0.2294,
+      "step": 21220
+    },
+    {
+      "epoch": 8.838951310861423,
+      "grad_norm": 0.2421875,
+      "learning_rate": 2.5538418952232054e-07,
+      "loss": 0.2324,
+      "step": 21240
+    },
+    {
+      "epoch": 8.847274240532668,
+      "grad_norm": 0.271484375,
+      "learning_rate": 2.5179841293197476e-07,
+      "loss": 0.2331,
+      "step": 21260
+    },
+    {
+      "epoch": 8.855597170203911,
+      "grad_norm": 0.25,
+      "learning_rate": 2.482366526960786e-07,
+      "loss": 0.2328,
+      "step": 21280
+    },
+    {
+      "epoch": 8.863920099875156,
+      "grad_norm": 0.28515625,
+      "learning_rate": 2.446989468630434e-07,
+      "loss": 0.2363,
+      "step": 21300
+    },
+    {
+      "epoch": 8.863920099875156,
+      "eval_main_loss": 0.23828859627246857,
+      "eval_main_runtime": 6.3239,
+      "eval_main_samples_per_second": 30.045,
+      "eval_main_steps_per_second": 3.795,
+      "step": 21300
+    },
+    {
+      "epoch": 8.863920099875156,
+      "eval_anatomy_loss": 2.8329224586486816,
+      "eval_anatomy_runtime": 0.267,
+      "eval_anatomy_samples_per_second": 7.491,
+      "eval_anatomy_steps_per_second": 3.746,
+      "step": 21300
+    },
+    {
+      "epoch": 8.863920099875156,
+      "eval_college_mathematics_loss": 2.0518035888671875,
+      "eval_college_mathematics_runtime": 0.2657,
+      "eval_college_mathematics_samples_per_second": 7.527,
+      "eval_college_mathematics_steps_per_second": 3.764,
+      "step": 21300
+    },
+    {
+      "epoch": 8.863920099875156,
+      "eval_international_law_loss": 3.0643723011016846,
+      "eval_international_law_runtime": 0.2657,
+      "eval_international_law_samples_per_second": 7.528,
+      "eval_international_law_steps_per_second": 3.764,
+      "step": 21300
+    },
+    {
+      "epoch": 8.8722430295464,
+      "grad_norm": 0.27734375,
+      "learning_rate": 2.411853332243183e-07,
+      "loss": 0.2284,
+      "step": 21320
+    },
+    {
+      "epoch": 8.880565959217645,
+      "grad_norm": 0.28515625,
+      "learning_rate": 2.376958493139886e-07,
+      "loss": 0.2363,
+      "step": 21340
+    },
+    {
+      "epoch": 8.88888888888889,
+      "grad_norm": 0.234375,
+      "learning_rate": 2.3423053240837518e-07,
+      "loss": 0.2325,
+      "step": 21360
+    },
+    {
+      "epoch": 8.897211818560134,
+      "grad_norm": 0.21875,
+      "learning_rate": 2.3078941952563466e-07,
+      "loss": 0.2328,
+      "step": 21380
+    },
+    {
+      "epoch": 8.905534748231377,
+      "grad_norm": 0.29296875,
+      "learning_rate": 2.2737254742536547e-07,
+      "loss": 0.23,
+      "step": 21400
+    },
+    {
+      "epoch": 8.905534748231377,
+      "eval_main_loss": 0.23842591047286987,
+      "eval_main_runtime": 6.3216,
+      "eval_main_samples_per_second": 30.056,
+      "eval_main_steps_per_second": 3.796,
+      "step": 21400
+    },
+    {
+      "epoch": 8.905534748231377,
+      "eval_anatomy_loss": 2.8358285427093506,
+      "eval_anatomy_runtime": 0.2663,
+      "eval_anatomy_samples_per_second": 7.509,
+      "eval_anatomy_steps_per_second": 3.755,
+      "step": 21400
+    },
+    {
+      "epoch": 8.905534748231377,
+      "eval_college_mathematics_loss": 2.0552380084991455,
+      "eval_college_mathematics_runtime": 0.2662,
+      "eval_college_mathematics_samples_per_second": 7.514,
+      "eval_college_mathematics_steps_per_second": 3.757,
+      "step": 21400
+    },
+    {
+      "epoch": 8.905534748231377,
+      "eval_international_law_loss": 3.0662267208099365,
+      "eval_international_law_runtime": 0.2673,
+      "eval_international_law_samples_per_second": 7.482,
+      "eval_international_law_steps_per_second": 3.741,
+      "step": 21400
+    },
+    {
+      "epoch": 8.913857677902621,
+      "grad_norm": 0.228515625,
+      "learning_rate": 2.2397995260821342e-07,
+      "loss": 0.2304,
+      "step": 21420
+    },
+    {
+      "epoch": 8.922180607573866,
+      "grad_norm": 0.326171875,
+      "learning_rate": 2.206116713154838e-07,
+      "loss": 0.2309,
+      "step": 21440
+    },
+    {
+      "epoch": 8.93050353724511,
+      "grad_norm": 0.263671875,
+      "learning_rate": 2.172677395287537e-07,
+      "loss": 0.2324,
+      "step": 21460
+    },
+    {
+      "epoch": 8.938826466916355,
+      "grad_norm": 0.3046875,
+      "learning_rate": 2.1394819296948616e-07,
+      "loss": 0.2306,
+      "step": 21480
+    },
+    {
+      "epoch": 8.947149396587598,
+      "grad_norm": 0.263671875,
+      "learning_rate": 2.106530670986498e-07,
+      "loss": 0.2323,
+      "step": 21500
+    },
+    {
+      "epoch": 8.947149396587598,
+      "eval_main_loss": 0.23836477100849152,
+      "eval_main_runtime": 6.3511,
+      "eval_main_samples_per_second": 29.916,
+      "eval_main_steps_per_second": 3.779,
+      "step": 21500
+    },
+    {
+      "epoch": 8.947149396587598,
+      "eval_anatomy_loss": 2.8358511924743652,
+      "eval_anatomy_runtime": 0.2678,
+      "eval_anatomy_samples_per_second": 7.468,
+      "eval_anatomy_steps_per_second": 3.734,
+      "step": 21500
+    },
+    {
+      "epoch": 8.947149396587598,
+      "eval_college_mathematics_loss": 2.0521907806396484,
+      "eval_college_mathematics_runtime": 0.2659,
+      "eval_college_mathematics_samples_per_second": 7.521,
+      "eval_college_mathematics_steps_per_second": 3.761,
+      "step": 21500
+    },
+    {
+      "epoch": 8.947149396587598,
+      "eval_international_law_loss": 3.0672295093536377,
+      "eval_international_law_runtime": 0.2671,
+      "eval_international_law_samples_per_second": 7.486,
+      "eval_international_law_steps_per_second": 3.743,
+      "step": 21500
+    },
+    {
+      "epoch": 8.955472326258842,
+      "grad_norm": 0.28515625,
+      "learning_rate": 2.0738239711634133e-07,
+      "loss": 0.2317,
+      "step": 21520
+    },
+    {
+      "epoch": 8.963795255930087,
+      "grad_norm": 0.2490234375,
+      "learning_rate": 2.0413621796140647e-07,
+      "loss": 0.2289,
+      "step": 21540
+    },
+    {
+      "epoch": 8.972118185601332,
+      "grad_norm": 0.28515625,
+      "learning_rate": 2.0091456431106854e-07,
+      "loss": 0.2357,
+      "step": 21560
+    },
+    {
+      "epoch": 8.980441115272576,
+      "grad_norm": 0.330078125,
+      "learning_rate": 1.977174705805582e-07,
+      "loss": 0.2299,
+      "step": 21580
+    },
+    {
+      "epoch": 8.98876404494382,
+      "grad_norm": 0.2353515625,
+      "learning_rate": 1.9454497092274565e-07,
+      "loss": 0.2327,
+      "step": 21600
+    },
+    {
+      "epoch": 8.98876404494382,
+      "eval_main_loss": 0.23839198052883148,
+      "eval_main_runtime": 6.3485,
+      "eval_main_samples_per_second": 29.928,
+      "eval_main_steps_per_second": 3.78,
+      "step": 21600
+    },
+    {
+      "epoch": 8.98876404494382,
+      "eval_anatomy_loss": 2.8339269161224365,
+      "eval_anatomy_runtime": 0.2668,
+      "eval_anatomy_samples_per_second": 7.495,
+      "eval_anatomy_steps_per_second": 3.747,
+      "step": 21600
+    },
+    {
+      "epoch": 8.98876404494382,
+      "eval_college_mathematics_loss": 2.0503244400024414,
+      "eval_college_mathematics_runtime": 0.2663,
+      "eval_college_mathematics_samples_per_second": 7.511,
+      "eval_college_mathematics_steps_per_second": 3.756,
+      "step": 21600
+    },
+    {
+      "epoch": 8.98876404494382,
+      "eval_international_law_loss": 3.066997766494751,
+      "eval_international_law_runtime": 0.2675,
+      "eval_international_law_samples_per_second": 7.477,
+      "eval_international_law_steps_per_second": 3.739,
+      "step": 21600
+    },
+    {
+      "epoch": 8.997086974615064,
+      "grad_norm": 0.3125,
+      "learning_rate": 1.9139709922777528e-07,
+      "loss": 0.234,
+      "step": 21620
+    },
+    {
+      "epoch": 9.005409904286308,
+      "grad_norm": 0.294921875,
+      "learning_rate": 1.8827388912270318e-07,
+      "loss": 0.2279,
+      "step": 21640
+    },
+    {
+      "epoch": 9.013732833957553,
+      "grad_norm": 0.3359375,
+      "learning_rate": 1.8517537397114066e-07,
+      "loss": 0.2288,
+      "step": 21660
+    },
+    {
+      "epoch": 9.022055763628797,
+      "grad_norm": 0.275390625,
+      "learning_rate": 1.8210158687289258e-07,
+      "loss": 0.2344,
+      "step": 21680
+    },
+    {
+      "epoch": 9.030378693300042,
+      "grad_norm": 0.267578125,
+      "learning_rate": 1.7905256066361037e-07,
+      "loss": 0.2313,
+      "step": 21700
+    },
+    {
+      "epoch": 9.030378693300042,
+      "eval_main_loss": 0.23835183680057526,
+      "eval_main_runtime": 6.3323,
+      "eval_main_samples_per_second": 30.005,
+      "eval_main_steps_per_second": 3.79,
+      "step": 21700
+    },
+    {
+      "epoch": 9.030378693300042,
+      "eval_anatomy_loss": 2.832937479019165,
+      "eval_anatomy_runtime": 0.2671,
+      "eval_anatomy_samples_per_second": 7.487,
+      "eval_anatomy_steps_per_second": 3.743,
+      "step": 21700
+    },
+    {
+      "epoch": 9.030378693300042,
+      "eval_college_mathematics_loss": 2.048555612564087,
+      "eval_college_mathematics_runtime": 0.2651,
+      "eval_college_mathematics_samples_per_second": 7.545,
+      "eval_college_mathematics_steps_per_second": 3.773,
+      "step": 21700
+    },
+    {
+      "epoch": 9.030378693300042,
+      "eval_international_law_loss": 3.0660414695739746,
+      "eval_international_law_runtime": 0.266,
+      "eval_international_law_samples_per_second": 7.519,
+      "eval_international_law_steps_per_second": 3.759,
+      "step": 21700
+    },
+    {
+      "epoch": 9.038701622971287,
+      "grad_norm": 0.255859375,
+      "learning_rate": 1.7602832791443648e-07,
+      "loss": 0.2293,
+      "step": 21720
+    },
+    {
+      "epoch": 9.04702455264253,
+      "grad_norm": 0.275390625,
+      "learning_rate": 1.7302892093165684e-07,
+      "loss": 0.2289,
+      "step": 21740
+    },
+    {
+      "epoch": 9.055347482313774,
+      "grad_norm": 0.328125,
+      "learning_rate": 1.700543717563591e-07,
+      "loss": 0.2313,
+      "step": 21760
+    },
+    {
+      "epoch": 9.063670411985019,
+      "grad_norm": 0.2890625,
+      "learning_rate": 1.6710471216408563e-07,
+      "loss": 0.2315,
+      "step": 21780
+    },
+    {
+      "epoch": 9.071993341656263,
+      "grad_norm": 0.2578125,
+      "learning_rate": 1.641799736644986e-07,
+      "loss": 0.2333,
+      "step": 21800
+    },
+    {
+      "epoch": 9.071993341656263,
+      "eval_main_loss": 0.2383333146572113,
+      "eval_main_runtime": 6.3262,
+      "eval_main_samples_per_second": 30.034,
+      "eval_main_steps_per_second": 3.794,
+      "step": 21800
+    },
+    {
+      "epoch": 9.071993341656263,
+      "eval_anatomy_loss": 2.831967353820801,
+      "eval_anatomy_runtime": 0.2665,
+      "eval_anatomy_samples_per_second": 7.504,
+      "eval_anatomy_steps_per_second": 3.752,
+      "step": 21800
+    },
+    {
+      "epoch": 9.071993341656263,
+      "eval_college_mathematics_loss": 2.050198554992676,
+      "eval_college_mathematics_runtime": 0.2663,
+      "eval_college_mathematics_samples_per_second": 7.51,
+      "eval_college_mathematics_steps_per_second": 3.755,
+      "step": 21800
+    },
+    {
+      "epoch": 9.071993341656263,
+      "eval_international_law_loss": 3.0654492378234863,
+      "eval_international_law_runtime": 0.2666,
+      "eval_international_law_samples_per_second": 7.503,
+      "eval_international_law_steps_per_second": 3.751,
+      "step": 21800
+    },
+    {
+      "epoch": 9.080316271327508,
+      "grad_norm": 0.333984375,
+      "learning_rate": 1.6128018750103975e-07,
+      "loss": 0.2335,
+      "step": 21820
+    },
+    {
+      "epoch": 9.088639200998752,
+      "grad_norm": 0.24609375,
+      "learning_rate": 1.5840538465059813e-07,
+      "loss": 0.2288,
+      "step": 21840
+    },
+    {
+      "epoch": 9.096962130669995,
+      "grad_norm": 0.26953125,
+      "learning_rate": 1.555555958231808e-07,
+      "loss": 0.2326,
+      "step": 21860
+    },
+    {
+      "epoch": 9.10528506034124,
+      "grad_norm": 0.263671875,
+      "learning_rate": 1.527308514615819e-07,
+      "loss": 0.2301,
+      "step": 21880
+    },
+    {
+      "epoch": 9.113607990012484,
+      "grad_norm": 0.31640625,
+      "learning_rate": 1.4993118174105842e-07,
+      "loss": 0.2275,
+      "step": 21900
+    },
+    {
+      "epoch": 9.113607990012484,
+      "eval_main_loss": 0.23825402557849884,
+      "eval_main_runtime": 6.3324,
+      "eval_main_samples_per_second": 30.005,
+      "eval_main_steps_per_second": 3.79,
+      "step": 21900
+    },
+    {
+      "epoch": 9.113607990012484,
+      "eval_anatomy_loss": 2.833402395248413,
+      "eval_anatomy_runtime": 0.2656,
+      "eval_anatomy_samples_per_second": 7.529,
+      "eval_anatomy_steps_per_second": 3.765,
+      "step": 21900
+    },
+    {
+      "epoch": 9.113607990012484,
+      "eval_college_mathematics_loss": 2.0520825386047363,
+      "eval_college_mathematics_runtime": 0.2664,
+      "eval_college_mathematics_samples_per_second": 7.508,
+      "eval_college_mathematics_steps_per_second": 3.754,
+      "step": 21900
+    },
+    {
+      "epoch": 9.113607990012484,
+      "eval_international_law_loss": 3.0643093585968018,
+      "eval_international_law_runtime": 0.2661,
+      "eval_international_law_samples_per_second": 7.515,
+      "eval_international_law_steps_per_second": 3.758,
+      "step": 21900
+    },
+    {
+      "epoch": 9.121930919683729,
+      "grad_norm": 0.275390625,
+      "learning_rate": 1.471566165690086e-07,
+      "loss": 0.2317,
+      "step": 21920
+    },
+    {
+      "epoch": 9.130253849354974,
+      "grad_norm": 0.25390625,
+      "learning_rate": 1.4440718558465294e-07,
+      "loss": 0.232,
+      "step": 21940
+    },
+    {
+      "epoch": 9.138576779026216,
+      "grad_norm": 0.27734375,
+      "learning_rate": 1.416829181587151e-07,
+      "loss": 0.2346,
+      "step": 21960
+    },
+    {
+      "epoch": 9.146899708697461,
+      "grad_norm": 0.212890625,
+      "learning_rate": 1.3898384339311038e-07,
+      "loss": 0.2318,
+      "step": 21980
+    },
+    {
+      "epoch": 9.155222638368706,
+      "grad_norm": 0.28125,
+      "learning_rate": 1.3630999012063467e-07,
+      "loss": 0.2328,
+      "step": 22000
+    },
+    {
+      "epoch": 9.155222638368706,
+      "eval_main_loss": 0.23840220272541046,
+      "eval_main_runtime": 6.333,
+      "eval_main_samples_per_second": 30.002,
+      "eval_main_steps_per_second": 3.79,
+      "step": 22000
+    },
+    {
+      "epoch": 9.155222638368706,
+      "eval_anatomy_loss": 2.8348910808563232,
+      "eval_anatomy_runtime": 0.2663,
+      "eval_anatomy_samples_per_second": 7.511,
+      "eval_anatomy_steps_per_second": 3.755,
+      "step": 22000
+    },
+    {
+      "epoch": 9.155222638368706,
+      "eval_college_mathematics_loss": 2.0505683422088623,
+      "eval_college_mathematics_runtime": 0.2666,
+      "eval_college_mathematics_samples_per_second": 7.501,
+      "eval_college_mathematics_steps_per_second": 3.75,
+      "step": 22000
+    },
+    {
+      "epoch": 9.155222638368706,
+      "eval_international_law_loss": 3.065016508102417,
+      "eval_international_law_runtime": 0.2668,
+      "eval_international_law_samples_per_second": 7.497,
+      "eval_international_law_steps_per_second": 3.748,
+      "step": 22000
+    },
+    {
+      "epoch": 9.16354556803995,
+      "grad_norm": 0.2734375,
+      "learning_rate": 1.3366138690465437e-07,
+      "loss": 0.2313,
+      "step": 22020
+    },
+    {
+      "epoch": 9.171868497711195,
+      "grad_norm": 0.294921875,
+      "learning_rate": 1.3103806203880493e-07,
+      "loss": 0.2352,
+      "step": 22040
+    },
+    {
+      "epoch": 9.18019142738244,
+      "grad_norm": 0.337890625,
+      "learning_rate": 1.2844004354668483e-07,
+      "loss": 0.2299,
+      "step": 22060
+    },
+    {
+      "epoch": 9.188514357053682,
+      "grad_norm": 0.3046875,
+      "learning_rate": 1.258673591815579e-07,
+      "loss": 0.2313,
+      "step": 22080
+    },
+    {
+      "epoch": 9.196837286724927,
+      "grad_norm": 0.267578125,
+      "learning_rate": 1.2332003642605868e-07,
+      "loss": 0.2304,
+      "step": 22100
+    },
+    {
+      "epoch": 9.196837286724927,
+      "eval_main_loss": 0.2382725030183792,
+      "eval_main_runtime": 6.3586,
+      "eval_main_samples_per_second": 29.881,
+      "eval_main_steps_per_second": 3.774,
+      "step": 22100
+    },
+    {
+      "epoch": 9.196837286724927,
+      "eval_anatomy_loss": 2.8336524963378906,
+      "eval_anatomy_runtime": 0.268,
+      "eval_anatomy_samples_per_second": 7.462,
+      "eval_anatomy_steps_per_second": 3.731,
+      "step": 22100
+    },
+    {
+      "epoch": 9.196837286724927,
+      "eval_college_mathematics_loss": 2.0524938106536865,
+      "eval_college_mathematics_runtime": 0.2678,
+      "eval_college_mathematics_samples_per_second": 7.468,
+      "eval_college_mathematics_steps_per_second": 3.734,
+      "step": 22100
+    },
+    {
+      "epoch": 9.196837286724927,
+      "eval_international_law_loss": 3.0638139247894287,
+      "eval_international_law_runtime": 0.2668,
+      "eval_international_law_samples_per_second": 7.495,
+      "eval_international_law_steps_per_second": 3.748,
+      "step": 22100
+    },
+    {
+      "epoch": 9.205160216396171,
+      "grad_norm": 0.29296875,
+      "learning_rate": 1.2079810249189415e-07,
+      "loss": 0.234,
+      "step": 22120
+    },
+    {
+      "epoch": 9.213483146067416,
+      "grad_norm": 0.2734375,
+      "learning_rate": 1.1830158431955841e-07,
+      "loss": 0.2319,
+      "step": 22140
+    },
+    {
+      "epoch": 9.22180607573866,
+      "grad_norm": 0.283203125,
+      "learning_rate": 1.1583050857804145e-07,
+      "loss": 0.2295,
+      "step": 22160
+    },
+    {
+      "epoch": 9.230129005409903,
+      "grad_norm": 0.29296875,
+      "learning_rate": 1.1338490166454386e-07,
+      "loss": 0.2315,
+      "step": 22180
+    },
+    {
+      "epoch": 9.238451935081148,
+      "grad_norm": 0.310546875,
+      "learning_rate": 1.109647897041985e-07,
+      "loss": 0.2329,
+      "step": 22200
+    },
+    {
+      "epoch": 9.238451935081148,
+      "eval_main_loss": 0.23832343518733978,
+      "eval_main_runtime": 6.3644,
+      "eval_main_samples_per_second": 29.854,
+      "eval_main_steps_per_second": 3.771,
+      "step": 22200
+    },
+    {
+      "epoch": 9.238451935081148,
+      "eval_anatomy_loss": 2.8315091133117676,
+      "eval_anatomy_runtime": 0.2675,
+      "eval_anatomy_samples_per_second": 7.477,
+      "eval_anatomy_steps_per_second": 3.739,
+      "step": 22200
+    },
+    {
+      "epoch": 9.238451935081148,
+      "eval_college_mathematics_loss": 2.051854133605957,
+      "eval_college_mathematics_runtime": 0.2658,
+      "eval_college_mathematics_samples_per_second": 7.525,
+      "eval_college_mathematics_steps_per_second": 3.763,
+      "step": 22200
+    },
+    {
+      "epoch": 9.238451935081148,
+      "eval_international_law_loss": 3.0672245025634766,
+      "eval_international_law_runtime": 0.2666,
+      "eval_international_law_samples_per_second": 7.501,
+      "eval_international_law_steps_per_second": 3.751,
+      "step": 22200
+    },
+    {
+      "epoch": 9.246774864752393,
+      "grad_norm": 0.283203125,
+      "learning_rate": 1.085701985497875e-07,
+      "loss": 0.2331,
+      "step": 22220
+    },
+    {
+      "epoch": 9.255097794423637,
+      "grad_norm": 0.2431640625,
+      "learning_rate": 1.062011537814675e-07,
+      "loss": 0.2341,
+      "step": 22240
+    },
+    {
+      "epoch": 9.263420724094882,
+      "grad_norm": 0.298828125,
+      "learning_rate": 1.0385768070649783e-07,
+      "loss": 0.2314,
+      "step": 22260
+    },
+    {
+      "epoch": 9.271743653766126,
+      "grad_norm": 0.31640625,
+      "learning_rate": 1.015398043589677e-07,
+      "loss": 0.2301,
+      "step": 22280
+    },
+    {
+      "epoch": 9.28006658343737,
+      "grad_norm": 0.2578125,
+      "learning_rate": 9.924754949953069e-08,
+      "loss": 0.2287,
+      "step": 22300
+    },
+    {
+      "epoch": 9.28006658343737,
+      "eval_main_loss": 0.23838910460472107,
+      "eval_main_runtime": 6.3361,
+      "eval_main_samples_per_second": 29.987,
+      "eval_main_steps_per_second": 3.788,
+      "step": 22300
+    },
+    {
+      "epoch": 9.28006658343737,
+      "eval_anatomy_loss": 2.8351857662200928,
+      "eval_anatomy_runtime": 0.2664,
+      "eval_anatomy_samples_per_second": 7.508,
+      "eval_anatomy_steps_per_second": 3.754,
+      "step": 22300
+    },
+    {
+      "epoch": 9.28006658343737,
+      "eval_college_mathematics_loss": 2.0533783435821533,
+      "eval_college_mathematics_runtime": 0.2663,
+      "eval_college_mathematics_samples_per_second": 7.51,
+      "eval_college_mathematics_steps_per_second": 3.755,
+      "step": 22300
+    },
+    {
+      "epoch": 9.28006658343737,
+      "eval_international_law_loss": 3.0666663646698,
+      "eval_international_law_runtime": 0.2658,
+      "eval_international_law_samples_per_second": 7.525,
+      "eval_international_law_steps_per_second": 3.762,
+      "step": 22300
+    },
+    {
+      "epoch": 9.288389513108614,
+      "grad_norm": 0.306640625,
+      "learning_rate": 9.698094061513868e-08,
+      "loss": 0.2336,
+      "step": 22320
+    },
+    {
+      "epoch": 9.296712442779858,
+      "grad_norm": 0.279296875,
+      "learning_rate": 9.474000191878163e-08,
+      "loss": 0.2304,
+      "step": 22340
+    },
+    {
+      "epoch": 9.305035372451103,
+      "grad_norm": 0.296875,
+      "learning_rate": 9.252475734922883e-08,
+      "loss": 0.2302,
+      "step": 22360
+    },
+    {
+      "epoch": 9.313358302122348,
+      "grad_norm": 0.31640625,
+      "learning_rate": 9.033523057077193e-08,
+      "loss": 0.2312,
+      "step": 22380
+    },
+    {
+      "epoch": 9.321681231793592,
+      "grad_norm": 0.30078125,
+      "learning_rate": 8.817144497297342e-08,
+      "loss": 0.2355,
+      "step": 22400
+    },
+    {
+      "epoch": 9.321681231793592,
+      "eval_main_loss": 0.23832879960536957,
+      "eval_main_runtime": 6.3273,
+      "eval_main_samples_per_second": 30.029,
+      "eval_main_steps_per_second": 3.793,
+      "step": 22400
+    },
+    {
+      "epoch": 9.321681231793592,
+      "eval_anatomy_loss": 2.8331458568573,
+      "eval_anatomy_runtime": 0.2668,
+      "eval_anatomy_samples_per_second": 7.497,
+      "eval_anatomy_steps_per_second": 3.748,
+      "step": 22400
+    },
+    {
+      "epoch": 9.321681231793592,
+      "eval_college_mathematics_loss": 2.0527069568634033,
+      "eval_college_mathematics_runtime": 0.2664,
+      "eval_college_mathematics_samples_per_second": 7.506,
+      "eval_college_mathematics_steps_per_second": 3.753,
+      "step": 22400
+    },
+    {
+      "epoch": 9.321681231793592,
+      "eval_international_law_loss": 3.0645482540130615,
+      "eval_international_law_runtime": 0.2664,
+      "eval_international_law_samples_per_second": 7.508,
+      "eval_international_law_steps_per_second": 3.754,
+      "step": 22400
+    },
+    {
+      "epoch": 9.330004161464835,
+      "grad_norm": 0.287109375,
+      "learning_rate": 8.603342367041578e-08,
+      "loss": 0.2334,
+      "step": 22420
+    },
+    {
+      "epoch": 9.33832709113608,
+      "grad_norm": 0.271484375,
+      "learning_rate": 8.392118950245581e-08,
+      "loss": 0.2309,
+      "step": 22440
+    },
+    {
+      "epoch": 9.346650020807324,
+      "grad_norm": 0.26953125,
+      "learning_rate": 8.183476503297982e-08,
+      "loss": 0.2301,
+      "step": 22460
+    },
+    {
+      "epoch": 9.354972950478569,
+      "grad_norm": 0.287109375,
+      "learning_rate": 7.977417255016162e-08,
+      "loss": 0.2335,
+      "step": 22480
+    },
+    {
+      "epoch": 9.363295880149813,
+      "grad_norm": 0.28125,
+      "learning_rate": 7.77394340662277e-08,
+      "loss": 0.2291,
+      "step": 22500
+    },
+    {
+      "epoch": 9.363295880149813,
+      "eval_main_loss": 0.23837444186210632,
+      "eval_main_runtime": 6.3339,
+      "eval_main_samples_per_second": 29.997,
+      "eval_main_steps_per_second": 3.789,
+      "step": 22500
+    },
+    {
+      "epoch": 9.363295880149813,
+      "eval_anatomy_loss": 2.834547758102417,
+      "eval_anatomy_runtime": 0.2678,
+      "eval_anatomy_samples_per_second": 7.467,
+      "eval_anatomy_steps_per_second": 3.734,
+      "step": 22500
+    },
+    {
+      "epoch": 9.363295880149813,
+      "eval_college_mathematics_loss": 2.0539867877960205,
+      "eval_college_mathematics_runtime": 0.2657,
+      "eval_college_mathematics_samples_per_second": 7.526,
+      "eval_college_mathematics_steps_per_second": 3.763,
+      "step": 22500
+    },
+    {
+      "epoch": 9.363295880149813,
+      "eval_international_law_loss": 3.0666635036468506,
+      "eval_international_law_runtime": 0.2663,
+      "eval_international_law_samples_per_second": 7.511,
+      "eval_international_law_steps_per_second": 3.756,
+      "step": 22500
+    },
+    {
+      "epoch": 9.371618809821056,
+      "grad_norm": 0.3125,
+      "learning_rate": 7.573057131721684e-08,
+      "loss": 0.23,
+      "step": 22520
+    },
+    {
+      "epoch": 9.3799417394923,
+      "grad_norm": 0.30859375,
+      "learning_rate": 7.374760576275397e-08,
+      "loss": 0.2319,
+      "step": 22540
+    },
+    {
+      "epoch": 9.388264669163545,
+      "grad_norm": 0.294921875,
+      "learning_rate": 7.179055858581586e-08,
+      "loss": 0.2296,
+      "step": 22560
+    },
+    {
+      "epoch": 9.39658759883479,
+      "grad_norm": 0.28515625,
+      "learning_rate": 6.985945069250766e-08,
+      "loss": 0.2351,
+      "step": 22580
+    },
+    {
+      "epoch": 9.404910528506035,
+      "grad_norm": 0.2734375,
+      "learning_rate": 6.795430271183929e-08,
+      "loss": 0.237,
+      "step": 22600
+    },
+    {
+      "epoch": 9.404910528506035,
+      "eval_main_loss": 0.23828138411045074,
+      "eval_main_runtime": 6.3303,
+      "eval_main_samples_per_second": 30.014,
+      "eval_main_steps_per_second": 3.791,
+      "step": 22600
+    },
+    {
+      "epoch": 9.404910528506035,
+      "eval_anatomy_loss": 2.8350436687469482,
+      "eval_anatomy_runtime": 0.2661,
+      "eval_anatomy_samples_per_second": 7.516,
+      "eval_anatomy_steps_per_second": 3.758,
+      "step": 22600
+    },
+    {
+      "epoch": 9.404910528506035,
+      "eval_college_mathematics_loss": 2.051068067550659,
+      "eval_college_mathematics_runtime": 0.2664,
+      "eval_college_mathematics_samples_per_second": 7.506,
+      "eval_college_mathematics_steps_per_second": 3.753,
+      "step": 22600
+    },
+    {
+      "epoch": 9.404910528506035,
+      "eval_international_law_loss": 3.065190315246582,
+      "eval_international_law_runtime": 0.2653,
+      "eval_international_law_samples_per_second": 7.537,
+      "eval_international_law_steps_per_second": 3.769,
+      "step": 22600
+    },
+    {
+      "epoch": 9.41323345817728,
+      "grad_norm": 0.291015625,
+      "learning_rate": 6.607513499550328e-08,
+      "loss": 0.233,
+      "step": 22620
+    },
+    {
+      "epoch": 9.421556387848522,
+      "grad_norm": 0.310546875,
+      "learning_rate": 6.422196761766031e-08,
+      "loss": 0.2309,
+      "step": 22640
+    },
+    {
+      "epoch": 9.429879317519767,
+      "grad_norm": 0.271484375,
+      "learning_rate": 6.239482037472156e-08,
+      "loss": 0.2313,
+      "step": 22660
+    },
+    {
+      "epoch": 9.438202247191011,
+      "grad_norm": 0.314453125,
+      "learning_rate": 6.059371278513942e-08,
+      "loss": 0.2308,
+      "step": 22680
+    },
+    {
+      "epoch": 9.446525176862256,
+      "grad_norm": 0.29296875,
+      "learning_rate": 5.881866408919912e-08,
+      "loss": 0.2296,
+      "step": 22700
+    },
+    {
+      "epoch": 9.446525176862256,
+      "eval_main_loss": 0.23832696676254272,
+      "eval_main_runtime": 6.3308,
+      "eval_main_samples_per_second": 30.012,
+      "eval_main_steps_per_second": 3.791,
+      "step": 22700
+    },
+    {
+      "epoch": 9.446525176862256,
+      "eval_anatomy_loss": 2.834747791290283,
+      "eval_anatomy_runtime": 0.2665,
+      "eval_anatomy_samples_per_second": 7.506,
+      "eval_anatomy_steps_per_second": 3.753,
+      "step": 22700
+    },
+    {
+      "epoch": 9.446525176862256,
+      "eval_college_mathematics_loss": 2.05255389213562,
+      "eval_college_mathematics_runtime": 0.2657,
+      "eval_college_mathematics_samples_per_second": 7.527,
+      "eval_college_mathematics_steps_per_second": 3.763,
+      "step": 22700
+    },
+    {
+      "epoch": 9.446525176862256,
+      "eval_international_law_loss": 3.064356803894043,
+      "eval_international_law_runtime": 0.266,
+      "eval_international_law_samples_per_second": 7.518,
+      "eval_international_law_steps_per_second": 3.759,
+      "step": 22700
+    },
+    {
+      "epoch": 9.4548481065335,
+      "grad_norm": 0.296875,
+      "learning_rate": 5.7069693248811566e-08,
+      "loss": 0.2266,
+      "step": 22720
+    },
+    {
+      "epoch": 9.463171036204745,
+      "grad_norm": 0.283203125,
+      "learning_rate": 5.5346818947311365e-08,
+      "loss": 0.2316,
+      "step": 22740
+    },
+    {
+      "epoch": 9.471493965875988,
+      "grad_norm": 0.265625,
+      "learning_rate": 5.3650059589258616e-08,
+      "loss": 0.2313,
+      "step": 22760
+    },
+    {
+      "epoch": 9.479816895547232,
+      "grad_norm": 0.26171875,
+      "learning_rate": 5.197943330024019e-08,
+      "loss": 0.2367,
+      "step": 22780
+    },
+    {
+      "epoch": 9.488139825218477,
+      "grad_norm": 0.32421875,
+      "learning_rate": 5.0334957926677917e-08,
+      "loss": 0.2302,
+      "step": 22800
+    },
+    {
+      "epoch": 9.488139825218477,
+      "eval_main_loss": 0.2384629100561142,
+      "eval_main_runtime": 6.3266,
+      "eval_main_samples_per_second": 30.032,
+      "eval_main_steps_per_second": 3.794,
+      "step": 22800
+    },
+    {
+      "epoch": 9.488139825218477,
+      "eval_anatomy_loss": 2.8347084522247314,
+      "eval_anatomy_runtime": 0.2653,
+      "eval_anatomy_samples_per_second": 7.54,
+      "eval_anatomy_steps_per_second": 3.77,
+      "step": 22800
+    },
+    {
+      "epoch": 9.488139825218477,
+      "eval_college_mathematics_loss": 2.0520851612091064,
+      "eval_college_mathematics_runtime": 0.267,
+      "eval_college_mathematics_samples_per_second": 7.492,
+      "eval_college_mathematics_steps_per_second": 3.746,
+      "step": 22800
+    },
+    {
+      "epoch": 9.488139825218477,
+      "eval_international_law_loss": 3.0670056343078613,
+      "eval_international_law_runtime": 0.2666,
+      "eval_international_law_samples_per_second": 7.501,
+      "eval_international_law_steps_per_second": 3.751,
+      "step": 22800
+    },
+    {
+      "epoch": 9.496462754889722,
+      "grad_norm": 0.26171875,
+      "learning_rate": 4.871665103563655e-08,
+      "loss": 0.2362,
+      "step": 22820
+    },
+    {
+      "epoch": 9.504785684560966,
+      "grad_norm": 0.298828125,
+      "learning_rate": 4.7124529914637226e-08,
+      "loss": 0.2338,
+      "step": 22840
+    },
+    {
+      "epoch": 9.513108614232209,
+      "grad_norm": 0.2353515625,
+      "learning_rate": 4.555861157147179e-08,
+      "loss": 0.2339,
+      "step": 22860
+    },
+    {
+      "epoch": 9.521431543903454,
+      "grad_norm": 0.259765625,
+      "learning_rate": 4.401891273402209e-08,
+      "loss": 0.2355,
+      "step": 22880
+    },
+    {
+      "epoch": 9.529754473574698,
+      "grad_norm": 0.26171875,
+      "learning_rate": 4.250544985008043e-08,
+      "loss": 0.2331,
+      "step": 22900
+    },
+    {
+      "epoch": 9.529754473574698,
+      "eval_main_loss": 0.2383948415517807,
+      "eval_main_runtime": 6.3315,
+      "eval_main_samples_per_second": 30.008,
+      "eval_main_steps_per_second": 3.791,
+      "step": 22900
+    },
+    {
+      "epoch": 9.529754473574698,
+      "eval_anatomy_loss": 2.832704782485962,
+      "eval_anatomy_runtime": 0.2655,
+      "eval_anatomy_samples_per_second": 7.532,
+      "eval_anatomy_steps_per_second": 3.766,
+      "step": 22900
+    },
+    {
+      "epoch": 9.529754473574698,
+      "eval_college_mathematics_loss": 2.0502614974975586,
+      "eval_college_mathematics_runtime": 0.2673,
+      "eval_college_mathematics_samples_per_second": 7.482,
+      "eval_college_mathematics_steps_per_second": 3.741,
+      "step": 22900
+    },
+    {
+      "epoch": 9.529754473574698,
+      "eval_international_law_loss": 3.0666775703430176,
+      "eval_international_law_runtime": 0.2657,
+      "eval_international_law_samples_per_second": 7.527,
+      "eval_international_law_steps_per_second": 3.764,
+      "step": 22900
+    },
+    {
+      "epoch": 9.538077403245943,
+      "grad_norm": 0.267578125,
+      "learning_rate": 4.1018239087174114e-08,
+      "loss": 0.2385,
+      "step": 22920
+    },
+    {
+      "epoch": 9.546400332917187,
+      "grad_norm": 0.2734375,
+      "learning_rate": 3.9557296332393413e-08,
+      "loss": 0.2306,
+      "step": 22940
+    },
+    {
+      "epoch": 9.554723262588432,
+      "grad_norm": 0.2353515625,
+      "learning_rate": 3.812263719222081e-08,
+      "loss": 0.2322,
+      "step": 22960
+    },
+    {
+      "epoch": 9.563046192259675,
+      "grad_norm": 0.2734375,
+      "learning_rate": 3.671427699236479e-08,
+      "loss": 0.2305,
+      "step": 22980
+    },
+    {
+      "epoch": 9.57136912193092,
+      "grad_norm": 0.30078125,
+      "learning_rate": 3.53322307775969e-08,
+      "loss": 0.2349,
+      "step": 23000
+    },
+    {
+      "epoch": 9.57136912193092,
+      "eval_main_loss": 0.23835638165473938,
+      "eval_main_runtime": 6.3283,
+      "eval_main_samples_per_second": 30.024,
+      "eval_main_steps_per_second": 3.792,
+      "step": 23000
+    },
+    {
+      "epoch": 9.57136912193092,
+      "eval_anatomy_loss": 2.8344595432281494,
+      "eval_anatomy_runtime": 0.2648,
+      "eval_anatomy_samples_per_second": 7.554,
+      "eval_anatomy_steps_per_second": 3.777,
+      "step": 23000
+    },
+    {
+      "epoch": 9.57136912193092,
+      "eval_college_mathematics_loss": 2.0526890754699707,
+      "eval_college_mathematics_runtime": 0.2671,
+      "eval_college_mathematics_samples_per_second": 7.488,
+      "eval_college_mathematics_steps_per_second": 3.744,
+      "step": 23000
+    },
+    {
+      "epoch": 9.57136912193092,
+      "eval_international_law_loss": 3.0659825801849365,
+      "eval_international_law_runtime": 0.2668,
+      "eval_international_law_samples_per_second": 7.496,
+      "eval_international_law_steps_per_second": 3.748,
+      "step": 23000
+    },
+    {
+      "epoch": 9.579692051602164,
+      "grad_norm": 0.24609375,
+      "learning_rate": 3.39765133115888e-08,
+      "loss": 0.2312,
+      "step": 23020
+    },
+    {
+      "epoch": 9.588014981273409,
+      "grad_norm": 0.28125,
+      "learning_rate": 3.264713907675687e-08,
+      "loss": 0.2308,
+      "step": 23040
+    },
+    {
+      "epoch": 9.596337910944653,
+      "grad_norm": 0.28515625,
+      "learning_rate": 3.134412227410677e-08,
+      "loss": 0.232,
+      "step": 23060
+    },
+    {
+      "epoch": 9.604660840615896,
+      "grad_norm": 0.3046875,
+      "learning_rate": 3.006747682308103e-08,
+      "loss": 0.2307,
+      "step": 23080
+    },
+    {
+      "epoch": 9.61298377028714,
+      "grad_norm": 0.291015625,
+      "learning_rate": 2.8817216361411438e-08,
+      "loss": 0.234,
+      "step": 23100
+    },
+    {
+      "epoch": 9.61298377028714,
+      "eval_main_loss": 0.23833413422107697,
+      "eval_main_runtime": 6.32,
+      "eval_main_samples_per_second": 30.063,
+      "eval_main_steps_per_second": 3.797,
+      "step": 23100
+    },
+    {
+      "epoch": 9.61298377028714,
+      "eval_anatomy_loss": 2.832688331604004,
+      "eval_anatomy_runtime": 0.2665,
+      "eval_anatomy_samples_per_second": 7.504,
+      "eval_anatomy_steps_per_second": 3.752,
+      "step": 23100
+    },
+    {
+      "epoch": 9.61298377028714,
+      "eval_college_mathematics_loss": 2.050745725631714,
+      "eval_college_mathematics_runtime": 0.2667,
+      "eval_college_mathematics_samples_per_second": 7.498,
+      "eval_college_mathematics_steps_per_second": 3.749,
+      "step": 23100
+    },
+    {
+      "epoch": 9.61298377028714,
+      "eval_international_law_loss": 3.0674684047698975,
+      "eval_international_law_runtime": 0.2656,
+      "eval_international_law_samples_per_second": 7.53,
+      "eval_international_law_steps_per_second": 3.765,
+      "step": 23100
+    },
+    {
+      "epoch": 9.621306699958385,
+      "grad_norm": 0.267578125,
+      "learning_rate": 2.7593354244972448e-08,
+      "loss": 0.2298,
+      "step": 23120
+    },
+    {
+      "epoch": 9.62962962962963,
+      "grad_norm": 0.298828125,
+      "learning_rate": 2.6395903547638825e-08,
+      "loss": 0.2305,
+      "step": 23140
+    },
+    {
+      "epoch": 9.637952559300874,
+      "grad_norm": 0.2578125,
+      "learning_rate": 2.5224877061146292e-08,
+      "loss": 0.2332,
+      "step": 23160
+    },
+    {
+      "epoch": 9.646275488972119,
+      "grad_norm": 0.287109375,
+      "learning_rate": 2.4080287294954706e-08,
+      "loss": 0.2306,
+      "step": 23180
+    },
+    {
+      "epoch": 9.654598418643362,
+      "grad_norm": 0.23828125,
+      "learning_rate": 2.2962146476114e-08,
+      "loss": 0.2315,
+      "step": 23200
+    },
+    {
+      "epoch": 9.654598418643362,
+      "eval_main_loss": 0.23836931586265564,
+      "eval_main_runtime": 6.3322,
+      "eval_main_samples_per_second": 30.005,
+      "eval_main_steps_per_second": 3.79,
+      "step": 23200
+    },
+    {
+      "epoch": 9.654598418643362,
+      "eval_anatomy_loss": 2.834285259246826,
+      "eval_anatomy_runtime": 0.2661,
+      "eval_anatomy_samples_per_second": 7.517,
+      "eval_anatomy_steps_per_second": 3.758,
+      "step": 23200
+    },
+    {
+      "epoch": 9.654598418643362,
+      "eval_college_mathematics_loss": 2.05277943611145,
+      "eval_college_mathematics_runtime": 0.2661,
+      "eval_college_mathematics_samples_per_second": 7.517,
+      "eval_college_mathematics_steps_per_second": 3.759,
+      "step": 23200
+    },
+    {
+      "epoch": 9.654598418643362,
+      "eval_international_law_loss": 3.065124273300171,
+      "eval_international_law_runtime": 0.2661,
+      "eval_international_law_samples_per_second": 7.517,
+      "eval_international_law_steps_per_second": 3.759,
+      "step": 23200
+    },
+    {
+      "epoch": 9.662921348314606,
+      "grad_norm": 0.26953125,
+      "learning_rate": 2.187046654913455e-08,
+      "loss": 0.2309,
+      "step": 23220
+    },
+    {
+      "epoch": 9.671244277985851,
+      "grad_norm": 0.26171875,
+      "learning_rate": 2.080525917585785e-08,
+      "loss": 0.2303,
+      "step": 23240
+    },
+    {
+      "epoch": 9.679567207657096,
+      "grad_norm": 0.2734375,
+      "learning_rate": 1.9766535735334102e-08,
+      "loss": 0.2316,
+      "step": 23260
+    },
+    {
+      "epoch": 9.68789013732834,
+      "grad_norm": 0.2734375,
+      "learning_rate": 1.875430732369954e-08,
+      "loss": 0.232,
+      "step": 23280
+    },
+    {
+      "epoch": 9.696213066999583,
+      "grad_norm": 0.294921875,
+      "learning_rate": 1.7768584754056796e-08,
+      "loss": 0.2298,
+      "step": 23300
+    },
+    {
+      "epoch": 9.696213066999583,
+      "eval_main_loss": 0.2383754402399063,
+      "eval_main_runtime": 6.3306,
+      "eval_main_samples_per_second": 30.013,
+      "eval_main_steps_per_second": 3.791,
+      "step": 23300
+    },
+    {
+      "epoch": 9.696213066999583,
+      "eval_anatomy_loss": 2.833543539047241,
+      "eval_anatomy_runtime": 0.2663,
+      "eval_anatomy_samples_per_second": 7.509,
+      "eval_anatomy_steps_per_second": 3.755,
+      "step": 23300
+    },
+    {
+      "epoch": 9.696213066999583,
+      "eval_college_mathematics_loss": 2.0523107051849365,
+      "eval_college_mathematics_runtime": 0.2657,
+      "eval_college_mathematics_samples_per_second": 7.527,
+      "eval_college_mathematics_steps_per_second": 3.763,
+      "step": 23300
+    },
+    {
+      "epoch": 9.696213066999583,
+      "eval_international_law_loss": 3.065718650817871,
+      "eval_international_law_runtime": 0.2663,
+      "eval_international_law_samples_per_second": 7.511,
+      "eval_international_law_steps_per_second": 3.756,
+      "step": 23300
+    },
+    {
+      "epoch": 9.704535996670828,
+      "grad_norm": 0.30859375,
+      "learning_rate": 1.6809378556361945e-08,
+      "loss": 0.2306,
+      "step": 23320
+    },
+    {
+      "epoch": 9.712858926342072,
+      "grad_norm": 0.30859375,
+      "learning_rate": 1.5876698977310145e-08,
+      "loss": 0.2311,
+      "step": 23340
+    },
+    {
+      "epoch": 9.721181856013317,
+      "grad_norm": 0.263671875,
+      "learning_rate": 1.497055598022601e-08,
+      "loss": 0.2291,
+      "step": 23360
+    },
+    {
+      "epoch": 9.729504785684561,
+      "grad_norm": 0.21875,
+      "learning_rate": 1.4090959244958402e-08,
+      "loss": 0.2285,
+      "step": 23380
+    },
+    {
+      "epoch": 9.737827715355806,
+      "grad_norm": 0.2890625,
+      "learning_rate": 1.3237918167776919e-08,
+      "loss": 0.2278,
+      "step": 23400
+    },
+    {
+      "epoch": 9.737827715355806,
+      "eval_main_loss": 0.23843945562839508,
+      "eval_main_runtime": 6.3292,
+      "eval_main_samples_per_second": 30.02,
+      "eval_main_steps_per_second": 3.792,
+      "step": 23400
+    },
+    {
+      "epoch": 9.737827715355806,
+      "eval_anatomy_loss": 2.8331258296966553,
+      "eval_anatomy_runtime": 0.2669,
+      "eval_anatomy_samples_per_second": 7.494,
+      "eval_anatomy_steps_per_second": 3.747,
+      "step": 23400
+    },
+    {
+      "epoch": 9.737827715355806,
+      "eval_college_mathematics_loss": 2.0540261268615723,
+      "eval_college_mathematics_runtime": 0.2653,
+      "eval_college_mathematics_samples_per_second": 7.539,
+      "eval_college_mathematics_steps_per_second": 3.769,
+      "step": 23400
+    },
+    {
+      "epoch": 9.737827715355806,
+      "eval_international_law_loss": 3.066096067428589,
+      "eval_international_law_runtime": 0.2663,
+      "eval_international_law_samples_per_second": 7.511,
+      "eval_international_law_steps_per_second": 3.755,
+      "step": 23400
+    },
+    {
+      "epoch": 9.746150645027049,
+      "grad_norm": 0.306640625,
+      "learning_rate": 1.2411441861269746e-08,
+      "loss": 0.2354,
+      "step": 23420
+    },
+    {
+      "epoch": 9.754473574698293,
+      "grad_norm": 0.251953125,
+      "learning_rate": 1.161153915424873e-08,
+      "loss": 0.2322,
+      "step": 23440
+    },
+    {
+      "epoch": 9.762796504369538,
+      "grad_norm": 0.224609375,
+      "learning_rate": 1.0838218591653348e-08,
+      "loss": 0.237,
+      "step": 23460
+    },
+    {
+      "epoch": 9.771119434040783,
+      "grad_norm": 0.3046875,
+      "learning_rate": 1.0091488434460506e-08,
+      "loss": 0.2304,
+      "step": 23480
+    },
+    {
+      "epoch": 9.779442363712027,
+      "grad_norm": 0.275390625,
+      "learning_rate": 9.371356659595431e-09,
+      "loss": 0.2284,
+      "step": 23500
+    },
+    {
+      "epoch": 9.779442363712027,
+      "eval_main_loss": 0.23842047154903412,
+      "eval_main_runtime": 6.3254,
+      "eval_main_samples_per_second": 30.038,
+      "eval_main_steps_per_second": 3.794,
+      "step": 23500
+    },
+    {
+      "epoch": 9.779442363712027,
+      "eval_anatomy_loss": 2.8333277702331543,
+      "eval_anatomy_runtime": 0.266,
+      "eval_anatomy_samples_per_second": 7.517,
+      "eval_anatomy_steps_per_second": 3.759,
+      "step": 23500
+    },
+    {
+      "epoch": 9.779442363712027,
+      "eval_college_mathematics_loss": 2.0542471408843994,
+      "eval_college_mathematics_runtime": 0.266,
+      "eval_college_mathematics_samples_per_second": 7.517,
+      "eval_college_mathematics_steps_per_second": 3.759,
+      "step": 23500
+    },
+    {
+      "epoch": 9.779442363712027,
+      "eval_international_law_loss": 3.0644288063049316,
+      "eval_international_law_runtime": 0.2657,
+      "eval_international_law_samples_per_second": 7.529,
+      "eval_international_law_steps_per_second": 3.764,
+      "step": 23500
+    },
+    {
+      "epoch": 9.787765293383272,
+      "grad_norm": 0.294921875,
+      "learning_rate": 8.677830959846756e-09,
+      "loss": 0.2291,
+      "step": 23520
+    },
+    {
+      "epoch": 9.796088223054515,
+      "grad_norm": 0.298828125,
+      "learning_rate": 8.010918743784624e-09,
+      "loss": 0.2362,
+      "step": 23540
+    },
+    {
+      "epoch": 9.80441115272576,
+      "grad_norm": 0.283203125,
+      "learning_rate": 7.370627135681319e-09,
+      "loss": 0.2299,
+      "step": 23560
+    },
+    {
+      "epoch": 9.812734082397004,
+      "grad_norm": 0.3046875,
+      "learning_rate": 6.75696297543521e-09,
+      "loss": 0.2291,
+      "step": 23580
+    },
+    {
+      "epoch": 9.821057012068248,
+      "grad_norm": 0.2890625,
+      "learning_rate": 6.169932818497482e-09,
+      "loss": 0.2299,
+      "step": 23600
+    },
+    {
+      "epoch": 9.821057012068248,
+      "eval_main_loss": 0.2384078949689865,
+      "eval_main_runtime": 6.3301,
+      "eval_main_samples_per_second": 30.015,
+      "eval_main_steps_per_second": 3.791,
+      "step": 23600
+    },
+    {
+      "epoch": 9.821057012068248,
+      "eval_anatomy_loss": 2.8333277702331543,
+      "eval_anatomy_runtime": 0.2668,
+      "eval_anatomy_samples_per_second": 7.495,
+      "eval_anatomy_steps_per_second": 3.748,
+      "step": 23600
+    },
+    {
+      "epoch": 9.821057012068248,
+      "eval_college_mathematics_loss": 2.0545294284820557,
+      "eval_college_mathematics_runtime": 0.2655,
+      "eval_college_mathematics_samples_per_second": 7.533,
+      "eval_college_mathematics_steps_per_second": 3.766,
+      "step": 23600
+    },
+    {
+      "epoch": 9.821057012068248,
+      "eval_international_law_loss": 3.0657763481140137,
+      "eval_international_law_runtime": 0.267,
+      "eval_international_law_samples_per_second": 7.492,
+      "eval_international_law_steps_per_second": 3.746,
+      "step": 23600
+    },
+    {
+      "epoch": 9.829379941739493,
+      "grad_norm": 0.314453125,
+      "learning_rate": 5.609542935802181e-09,
+      "loss": 0.2375,
+      "step": 23620
+    },
+    {
+      "epoch": 9.837702871410738,
+      "grad_norm": 0.236328125,
+      "learning_rate": 5.075799313699892e-09,
+      "loss": 0.2332,
+      "step": 23640
+    },
+    {
+      "epoch": 9.84602580108198,
+      "grad_norm": 0.3359375,
+      "learning_rate": 4.568707653892779e-09,
+      "loss": 0.2306,
+      "step": 23660
+    },
+    {
+      "epoch": 9.854348730753225,
+      "grad_norm": 0.29296875,
+      "learning_rate": 4.088273373373807e-09,
+      "loss": 0.2353,
+      "step": 23680
+    },
+    {
+      "epoch": 9.86267166042447,
+      "grad_norm": 0.32421875,
+      "learning_rate": 3.634501604370122e-09,
+      "loss": 0.2306,
+      "step": 23700
+    },
+    {
+      "epoch": 9.86267166042447,
+      "eval_main_loss": 0.23842915892601013,
+      "eval_main_runtime": 6.3292,
+      "eval_main_samples_per_second": 30.019,
+      "eval_main_steps_per_second": 3.792,
+      "step": 23700
+    },
+    {
+      "epoch": 9.86267166042447,
+      "eval_anatomy_loss": 2.8333277702331543,
+      "eval_anatomy_runtime": 0.2678,
+      "eval_anatomy_samples_per_second": 7.467,
+      "eval_anatomy_steps_per_second": 3.734,
+      "step": 23700
+    },
+    {
+      "epoch": 9.86267166042447,
+      "eval_college_mathematics_loss": 2.0545294284820557,
+      "eval_college_mathematics_runtime": 0.2668,
+      "eval_college_mathematics_samples_per_second": 7.495,
+      "eval_college_mathematics_steps_per_second": 3.747,
+      "step": 23700
+    },
+    {
+      "epoch": 9.86267166042447,
+      "eval_international_law_loss": 3.066246271133423,
+      "eval_international_law_runtime": 0.2662,
+      "eval_international_law_samples_per_second": 7.514,
+      "eval_international_law_steps_per_second": 3.757,
+      "step": 23700
+    },
+    {
+      "epoch": 9.870994590095714,
+      "grad_norm": 0.314453125,
+      "learning_rate": 3.2073971942864214e-09,
+      "loss": 0.2292,
+      "step": 23720
+    },
+    {
+      "epoch": 9.879317519766959,
+      "grad_norm": 0.287109375,
+      "learning_rate": 2.806964705654447e-09,
+      "loss": 0.2297,
+      "step": 23740
+    },
+    {
+      "epoch": 9.887640449438202,
+      "grad_norm": 0.26953125,
+      "learning_rate": 2.4332084160835766e-09,
+      "loss": 0.2313,
+      "step": 23760
+    },
+    {
+      "epoch": 9.895963379109446,
+      "grad_norm": 0.279296875,
+      "learning_rate": 2.086132318215861e-09,
+      "loss": 0.2343,
+      "step": 23780
+    },
+    {
+      "epoch": 9.90428630878069,
+      "grad_norm": 0.279296875,
+      "learning_rate": 1.765740119682169e-09,
+      "loss": 0.2333,
+      "step": 23800
+    },
+    {
+      "epoch": 9.90428630878069,
+      "eval_main_loss": 0.23843587934970856,
+      "eval_main_runtime": 6.3325,
+      "eval_main_samples_per_second": 30.004,
+      "eval_main_steps_per_second": 3.79,
+      "step": 23800
+    },
+    {
+      "epoch": 9.90428630878069,
+      "eval_anatomy_loss": 2.8333277702331543,
+      "eval_anatomy_runtime": 0.2662,
+      "eval_anatomy_samples_per_second": 7.514,
+      "eval_anatomy_steps_per_second": 3.757,
+      "step": 23800
+    },
+    {
+      "epoch": 9.90428630878069,
+      "eval_college_mathematics_loss": 2.0545294284820557,
+      "eval_college_mathematics_runtime": 0.2677,
+      "eval_college_mathematics_samples_per_second": 7.471,
+      "eval_college_mathematics_steps_per_second": 3.736,
+      "step": 23800
+    },
+    {
+      "epoch": 9.90428630878069,
+      "eval_international_law_loss": 3.066246271133423,
+      "eval_international_law_runtime": 0.2655,
+      "eval_international_law_samples_per_second": 7.534,
+      "eval_international_law_steps_per_second": 3.767,
+      "step": 23800
+    },
+    {
+      "epoch": 9.912609238451935,
+      "grad_norm": 0.28515625,
+      "learning_rate": 1.4720352430644402e-09,
+      "loss": 0.2361,
+      "step": 23820
+    },
+    {
+      "epoch": 9.92093216812318,
+      "grad_norm": 0.302734375,
+      "learning_rate": 1.205020825856551e-09,
+      "loss": 0.2316,
+      "step": 23840
+    },
+    {
+      "epoch": 9.929255097794425,
+      "grad_norm": 0.279296875,
+      "learning_rate": 9.64699720433504e-10,
+      "loss": 0.2324,
+      "step": 23860
+    },
+    {
+      "epoch": 9.937578027465667,
+      "grad_norm": 0.26953125,
+      "learning_rate": 7.510744940192327e-10,
+      "loss": 0.2323,
+      "step": 23880
+    },
+    {
+      "epoch": 9.945900957136912,
+      "grad_norm": 0.32421875,
+      "learning_rate": 5.64147428659123e-10,
+      "loss": 0.2323,
+      "step": 23900
+    },
+    {
+      "epoch": 9.945900957136912,
+      "eval_main_loss": 0.23843084275722504,
+      "eval_main_runtime": 6.3373,
+      "eval_main_samples_per_second": 29.981,
+      "eval_main_steps_per_second": 3.787,
+      "step": 23900
+    },
+    {
+      "epoch": 9.945900957136912,
+      "eval_anatomy_loss": 2.8333277702331543,
+      "eval_anatomy_runtime": 0.2674,
+      "eval_anatomy_samples_per_second": 7.479,
+      "eval_anatomy_steps_per_second": 3.74,
+      "step": 23900
+    },
+    {
+      "epoch": 9.945900957136912,
+      "eval_college_mathematics_loss": 2.0545294284820557,
+      "eval_college_mathematics_runtime": 0.2668,
+      "eval_college_mathematics_samples_per_second": 7.496,
+      "eval_college_mathematics_steps_per_second": 3.748,
+      "step": 23900
+    },
+    {
+      "epoch": 9.945900957136912,
+      "eval_international_law_loss": 3.066246271133423,
+      "eval_international_law_runtime": 0.2661,
+      "eval_international_law_samples_per_second": 7.517,
+      "eval_international_law_steps_per_second": 3.759,
+      "step": 23900
+    },
+    {
+      "epoch": 9.954223886808157,
+      "grad_norm": 0.310546875,
+      "learning_rate": 4.0392052119697654e-10,
+      "loss": 0.2338,
+      "step": 23920
+    },
+    {
+      "epoch": 9.962546816479401,
+      "grad_norm": 0.3125,
+      "learning_rate": 2.703954832528055e-10,
+      "loss": 0.2311,
+      "step": 23940
+    },
+    {
+      "epoch": 9.970869746150646,
+      "grad_norm": 0.2421875,
+      "learning_rate": 1.6357374120368197e-10,
+      "loss": 0.2312,
+      "step": 23960
+    },
+    {
+      "epoch": 9.979192675821889,
+      "grad_norm": 0.310546875,
+      "learning_rate": 8.345643617069243e-11,
+      "loss": 0.232,
+      "step": 23980
+    },
+    {
+      "epoch": 9.987515605493133,
+      "grad_norm": 0.240234375,
+      "learning_rate": 3.004442400422747e-11,
+      "loss": 0.2373,
+      "step": 24000
+    },
+    {
+      "epoch": 9.987515605493133,
+      "eval_main_loss": 0.23843084275722504,
+      "eval_main_runtime": 6.3378,
+      "eval_main_samples_per_second": 29.979,
+      "eval_main_steps_per_second": 3.787,
+      "step": 24000
+    },
+    {
+      "epoch": 9.987515605493133,
+      "eval_anatomy_loss": 2.8333277702331543,
+      "eval_anatomy_runtime": 0.2659,
+      "eval_anatomy_samples_per_second": 7.522,
+      "eval_anatomy_steps_per_second": 3.761,
+      "step": 24000
+    },
+    {
+      "epoch": 9.987515605493133,
+      "eval_college_mathematics_loss": 2.0545294284820557,
+      "eval_college_mathematics_runtime": 0.2659,
+      "eval_college_mathematics_samples_per_second": 7.521,
+      "eval_college_mathematics_steps_per_second": 3.761,
+      "step": 24000
+    },
+    {
+      "epoch": 9.987515605493133,
+      "eval_international_law_loss": 3.066246271133423,
+      "eval_international_law_runtime": 0.2672,
+      "eval_international_law_samples_per_second": 7.486,
+      "eval_international_law_steps_per_second": 3.743,
+      "step": 24000
+    },
+    {
+      "epoch": 9.995838535164378,
+      "grad_norm": 0.259765625,
+      "learning_rate": 3.3382752773203352e-12,
+      "loss": 0.2333,
+      "step": 24020
+    }
+  ],
+  "logging_steps": 20,
+  "max_steps": 24030,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.7646735410967085e+19,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}