ShyliaSafetensors/EnceladusHyperStock-24B
EnceladusHyperStock-24B by ShyliaSafetensors is an experimental 24 billion parameter multimerge model designed for complex scene handling, story-writing, and role-play. It is specifically optimized for creativity and consistency in long-form uncensored content. This model incorporates advanced merge methods like Model Stock and Dare-Ties, and features a unique `` tag for visible reasoning, making it suitable for intricate instruction following and conversational AI with personality depth.
Loading preview...
EnceladusHyperStock-24B: An Experimental Multimerge Model
EnceladusHyperStock-24B is an experimental 24 billion parameter language model developed by ShyliaSafetensors. It is a multimerge of several pre-trained models, including Dolphin-Mistral-GLM-4.7-Flash-24B-Venice-Edition-Thinking-Uncensored, Sakura-24B-Spice, and Cydonia-24B-Heretic-v4, created using mergekit with advanced techniques like Model Stock and Dare-Ties.
Key Capabilities & Features
- Enhanced Creativity and Consistency: Specifically designed to handle complex scenes, story-writing, and role-play with a focus on creative and consistent output.
- Visible Reasoning: Utilizes a unique
<thinking>tag mechanism, allowing the model to display its thought process for complex instruction following. - Uncensored Content Generation: Intended for long-form uncensored roleplay, creative writing, and conversational AI with deep personality.
Intended Use Cases
- Long-form uncensored roleplay and collaborative storytelling.
- Uncensored creative writing.
- Complex instruction following, especially where visible reasoning is beneficial.
- Developing conversational AI with nuanced personality.
This model is provided as an experimental version, building upon previous multimerge efforts, and is recommended for users seeking highly creative and consistent outputs in uncensored, long-form text generation.