flammenai/flammen15-gutenberg-DPO-v1-7B
flammenai/flammen15-gutenberg-DPO-v1-7B is a 7 billion parameter Mistral-based LLM, fine-tuned by flammenai using Direct Preference Optimization on Jon Durbin's Gutenberg DPO dataset. This model specializes in exceptional character roleplay, creative writing, and general intelligence, offering a 4096-token context length. It is designed for applications requiring nuanced conversational abilities and imaginative text generation.
Loading preview...
Overview
flammen15-gutenberg-DPO-v1-7B is a 7 billion parameter language model built upon the Mistral architecture. Developed by flammenai, this model was created by merging pretrained models and subsequently fine-tuned using Direct Preference Optimization (DPO) on Jon Durbin's Gutenberg DPO dataset. The fine-tuning process utilized an A100 GPU on Google Colab, following a methodology detailed in a guide by Maxime Labonne on fine-tuning Mistral-7B with DPO.
Key Capabilities
- Exceptional Character Roleplay: Optimized for generating responses that maintain consistent character personas and dialogue styles.
- Creative Writing: Excels at producing imaginative and coherent long-form text, suitable for storytelling and content generation.
- General Intelligence: Demonstrates strong performance across a range of general language understanding and generation tasks.
Performance Metrics
Evaluations on the Open LLM Leaderboard show an average score of 21.46. Notable scores include 47.98 on IFEval (0-Shot) and 32.67 on BBH (3-Shot), indicating its strengths in instruction following and complex reasoning, respectively. Detailed results are available on the Open LLM Leaderboard.
Good for
- Developing AI companions or chatbots requiring deep character immersion.
- Generating creative content such as stories, scripts, or descriptive passages.
- Applications needing a versatile language model with a focus on nuanced text generation.