Llama DNA 1.0 8B Instruct

🚀 DNA 1.0 8B Instruct

DNA 1.0 8B Instruct is a state - of - the - art (SOTA) bilingual language model based on the Llama architecture. It is specifically optimized for Korean language understanding and generation while maintaining strong English capabilities. This model was developed through a complex process, including model merging via spherical linear interpolation (SLERP) with Llama 3.1 8B Instruct, knowledge distillation (KD) using Llama 3.1 405B as the teacher model, continual pre - training (CPT) with a high - quality Korean dataset, and supervised fine - tuning (SFT) and direct preference optimization (DPO) to align with human preferences and enhance instruction - following abilities.

🚀 Quick Start

This model requires transformers >= 4.43.0.

Basic Usage

from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer

tokenizer = AutoTokenizer.from_pretrained('dnotitia/Llama-DNA-1.0-8B-Instruct')
model = AutoModelForCausalLM.from_pretrained('dnotitia/Llama-DNA-1.0-8B-Instruct', device_map='auto')
streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)

conversation = [
    {"role": "system", "content": "You are a helpful assistant, Dnotitia DNA."},
    {"role": "user", "content": "너의 이름은?"},
]
inputs = tokenizer.apply_chat_template(conversation,
                                       add_generation_prompt=True,
                                       return_dict=True,
                                       return_tensors="pt").to(model.device)
_ = model.generate(**inputs, streamer=streamer)

✨ Features

• Bilingual Excellence: Optimized for Korean while maintaining strong English capabilities.
• Sophisticated Development Process: Involves model merging, knowledge distillation, continual pre - training, supervised fine - tuning, and direct preference optimization.
• Extensive Instruction Tuning: Fine - tuned on approximately 7B tokens of curated data to follow complex instructions and engage in natural conversations.

📦 Information

📚 Documentation

Evaluation

We evaluated DNA 1.0 8B Instruct against other prominent language models of similar size across various benchmarks, including Korean - specific tasks and general language understanding metrics.

• The highest scores are in bold form, and the second - highest scores are underlined.

Evaluation Protocol
For easy reproduction of our evaluation results, we list the evaluation tools and settings used below:

Limitations

While DNA 1.0 8B Instruct demonstrates strong performance, users should be aware of the following limitations:

• The model may occasionally generate biased or inappropriate content.
• Responses are based on training data and may not reflect current information.
• The model may sometimes produce factually incorrect or inconsistent answers.
• Performance may vary depending on the complexity and domain of the task.
• Generated content should be reviewed for accuracy and appropriateness.

Appendix

• KMMLU scores comparison chart:

• DNA 1.0 8B Instruct model architecture ¹:

• The median percentage of model’s weight difference between before and after the merge (our SFT model + Llama 3.1 8B Instruct):

📄 License

This model is released under CC BY - NC 4.0 license. For commercial usage inquiries, please Contact us.

📖 Citation

If you use or discuss this model in your academic research, please cite the project to help spread awareness:

@misc{lee2025dna10technicalreport,
      title={DNA 1.0 Technical Report}, 
      author={Jungyup Lee and Jemin Kim and Sang Park and SeungJae Lee},
      year={2025},
      eprint={2501.10648},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2501.10648}, 
}

⚠️ Important Note

The model may generate biased, inappropriate, factually incorrect or inconsistent content. Generated content should be reviewed for accuracy and appropriateness.

💡 Usage Tip

For commercial usage of the model, please Contact us through the provided link.