LidongBing commited on
Commit
b1fcc9e
1 Parent(s): 5fbe396

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +2 -2
README.md CHANGED
@@ -14,7 +14,7 @@ tags:
14
  - multilingual
15
  ---
16
 
17
- # *SeaLLM3* - Large Language Models for Southeast Asia
18
 
19
 
20
  <p align="center">
@@ -29,7 +29,7 @@ tags:
29
  <a href="https://arxiv.org/pdf/2312.00738.pdf" target="_blank" rel="noopener">Technical Report</a>
30
  </p>
31
 
32
- We introduce **SeaLLM3**, the latest series of the SeaLLMs (Large Language Models for Southeast Asian languages) family. It achieves state-of-the-art performance among models with similar sizes, excelling across a diverse array of tasks such as world knowledge, mathematical reasoning, translation, and instruction following. In the meantime, it was specifically enhanced to be more trustworthy, exhibiting reduced hallucination and providing safe responses, particularly in queries closed related to Southeast Asian culture.
33
 
34
  ## 🔥 Highlights
35
  - State-of-the-art performance compared to open-source models of similar sizes, evaluated across various dimensions such as human exam questions, instruction-following, mathematics, and translation.
 
14
  - multilingual
15
  ---
16
 
17
+ # *SeaLLMs-v3* - Large Language Models for Southeast Asia
18
 
19
 
20
  <p align="center">
 
29
  <a href="https://arxiv.org/pdf/2312.00738.pdf" target="_blank" rel="noopener">Technical Report</a>
30
  </p>
31
 
32
+ We introduce **SeaLLMs-v3**, the latest series of the SeaLLMs (Large Language Models for Southeast Asian languages) family. It achieves state-of-the-art performance among models with similar sizes, excelling across a diverse array of tasks such as world knowledge, mathematical reasoning, translation, and instruction following. In the meantime, it was specifically enhanced to be more trustworthy, exhibiting reduced hallucination and providing safe responses, particularly in queries closed related to Southeast Asian culture.
33
 
34
  ## 🔥 Highlights
35
  - State-of-the-art performance compared to open-source models of similar sizes, evaluated across various dimensions such as human exam questions, instruction-following, mathematics, and translation.