metadata

license: apache-2.0
tags:
  - generated_from_trainer
base_model: yanolja/EEVE-Korean-2.8B-v1.0
model-index:
  - name: yanolja/EEVE-Korean-Instruct-2.8B-v1.0
    results: []

EEVE-Korean-Instruct-2.8B-v1.0

Join Our Community on Discord!

If you're passionate about the field of Large Language Models and wish to exchange knowledge and insights, we warmly invite you to join our Discord server. It's worth noting that Korean is the primary language used in this server. The landscape of LLM is evolving rapidly, and without active sharing, our collective knowledge risks becoming outdated swiftly. Let's collaborate and drive greater impact together! Join us here: Discord Link.

Our Dedicated Team (Alphabetical Order)

Research	Engineering	Product Management	UX Design
Myeongho Jeong	Geon Kim	Bokyung Huh	Eunsue Choi
Seungduk Kim	Rifqi Alfi
Seungtaek Choi	Sanghoon Han
	Suhyun Kang

About the Model

This model is a fine-tuned version of yanolja/EEVE-Korean-2.8B-v1.0, which is a Korean vocabulary-extended version of microsoft/phi-2. Specifically, we utilized Direct Preference Optimization (DPO) through the use of Axolotl.

For more details, please refer to our technical report: Efficient and Effective Vocabulary Expansion Towards Multilingual Large Language Models.

Prompt Template

A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions.
Human: {prompt}
Assistant:

How to Use it

from transformers import AutoTokenizer
from transformers import AutoModelForCausalLM

model = AutoModelForCausalLM.from_pretrained("yanolja/EEVE-Korean-Instruct-2.8B-v1.0", trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained("yanolja/EEVE-Korean-Instruct-2.8B-v1.0", trust_remote_code=True)

prompt_template = "A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions.\nHuman: {prompt}\nAssistant:\n"
text = '한국의 수도는 어디인가요? 아래 선택지 중 골라주세요.\n\n(A) 경성\n(B) 부산\n(C) 평양\n(D) 서울\n(E) 전주'
model_inputs = tokenizer(prompt_template.format(prompt=text), return_tensors='pt')

outputs = model.generate(**model_inputs, max_new_tokens=256)
output_text = tokenizer.batch_decode(outputs, skip_special_tokens=True)[0]
print(output_text)

Example Output

A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions.
Human: 한국의 수도는 어디인가요? 아래 선택지 중 골라주세요.

(A) 경성
(B) 부산
(C) 평양
(D) 서울
(E) 전주
Assistant:
한국의 수도는 (D) 서울입니다. 서울은 수도권과 수도권 내의 주요 도시들을 포함하는 광역 행정구역으로, 대한민국의 수도입니다. 서울은 수도권 인구의 약 70%를 차지하며, 대한민국의 경제, 정치, 문화의 중심지입니다.

Training Data

Korean-translated version of Open-Orca/SlimOrca-Dedup
Korean-translated version of argilla/ultrafeedback-binarized-preferences-cleaned
No other dataset was used

Citation

@misc{kim2024efficient,
      title={Efficient and Effective Vocabulary Expansion Towards Multilingual Large Language Models}, 
      author={Seungduk Kim and Seungtaek Choi and Myeongho Jeong},
      year={2024},
      eprint={2402.14714},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

@misc{cui2023ultrafeedback,
      title={UltraFeedback: Boosting Language Models with High-quality Feedback}, 
      author={Ganqu Cui and Lifan Yuan and Ning Ding and Guanming Yao and Wei Zhu and Yuan Ni and Guotong Xie and Zhiyuan Liu and Maosong Sun},
      year={2023},
      eprint={2310.01377},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

@misc{SlimOrcaDedup,
  title = {SlimOrca Dedup: A Deduplicated Subset of SlimOrca},
  author = {Wing Lian and Guan Wang and Bleys Goodson and Eugene Pentland and Austin Cook and Chanvichet Vong and "Teknium" and Nathan Hoos},
  year = {2023},
  publisher = {HuggingFace},
  url = {https://huggingface.co/datasets/Open-Orca/SlimOrca-Dedup/}
}

@misc{mukherjee2023orca,
      title={Orca: Progressive Learning from Complex Explanation Traces of GPT-4}, 
      author={Subhabrata Mukherjee and Arindam Mitra and Ganesh Jawahar and Sahaj Agarwal and Hamid Palangi and Ahmed Awadallah},
      year={2023},
      eprint={2306.02707},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	58.71
AI2 Reasoning Challenge (25-Shot)	58.28
HellaSwag (10-Shot)	72.42
MMLU (5-Shot)	53.35
TruthfulQA (0-shot)	48.32
Winogrande (5-shot)	74.82
GSM8k (5-shot)	45.11