922-CA/Llama-3-natsuki-ddlc-8b-v1
922-CA/Llama-3-natsuki-ddlc-8b-v1 is an 8 billion parameter Llama-3 based language model fine-tuned to embody the character Natsuki from Doki Doki Literature Club. Developed by 922-CA, this model specializes in generating responses consistent with Natsuki's personality and dialogue patterns. It is optimized for character-specific roleplay and interactive storytelling within the DDLC universe, offering an 8192 token context length. The model was trained using Unsloth and Huggingface's TRL library for efficient fine-tuning.
Loading preview...
Model Overview
922-CA/Llama-3-natsuki-ddlc-8b-v1 is an 8 billion parameter model built upon the Llama-3 architecture, specifically fine-tuned to emulate the character Natsuki from Doki Doki Literature Club. This model aims to provide character-accurate responses and dialogue, making it suitable for interactive experiences centered around Natsuki.
Key Characteristics
- Character Emulation: Designed to closely reflect Natsuki's personality and speaking style.
- Base Model: Utilizes the robust Meta-Llama-3-8B as its foundation.
- Efficient Training: Fine-tuned using Unsloth and Huggingface's TRL library, enabling faster training.
- Context Length: Supports an 8192 token context window.
Usage Recommendations
For optimal performance and character consistency, users are advised to format prompts by replacing "Human" and "Assistant" with "Player" and "Natsuki" respectively, as follows:
Player: (prompt)
Natsuki:Performance Metrics
Evaluated on the Open LLM Leaderboard, the model achieved an average score of 62.02. Notable scores include 80.99 on HellaSwag (10-Shot) and 64.75 on MMLU (5-Shot). Detailed evaluation results are available on the Hugging Face Open LLM Leaderboard.
Important Considerations
While fine-tuned for Natsuki's characteristics, the model may occasionally hallucinate or act out of character. Users should also be aware that the model is not guaranteed to produce aligned or safe outputs, and usage is at one's own risk.