flammenai/flammen17-mistral-7B
flammenai/flammen17-mistral-7B is a 7 billion parameter Mistral-based large language model, created by flammenai through merging and finetuning pretrained models. This model specializes in exceptional character roleplay, creative writing, and general intelligence. It leverages a 4096-token context length, making it suitable for nuanced and extended generative tasks. Its unique merge strategy aims to enhance specific creative and roleplaying capabilities.
Loading preview...
Overview
flammenai/flammen17-mistral-7B is a 7 billion parameter large language model built upon the Mistral architecture. It was developed by flammenai using a merge of pretrained models, specifically combining nbeerbower/Flammen-Bophades-7B and nbeerbower/flammen16-mistral-7B through the SLERP merge method. This approach aims to synthesize the strengths of its constituent models.
Key Capabilities
- Exceptional Character Roleplay: Designed to excel in generating consistent and engaging character interactions.
- Creative Writing: Optimized for producing high-quality, imaginative text across various creative writing formats.
- General Intelligence: Demonstrates strong performance in broader language understanding and generation tasks.
Merge Details
The model's unique characteristics stem from its specific merge configuration. The SLERP (Spherical Linear Interpolation) method was applied, with particular layer ranges and parameter weightings (t values) defined for self-attention and MLP components, indicating a fine-tuned balance between the merged models' contributions. This precise merging strategy is central to its specialized performance in creative and roleplaying applications.