Naphula/Cyberdonia-24B-v8
Naphula/Cyberdonia-24B-v8 is a 24 billion parameter language model created by Naphula using the `della` merge method with mergekit-exp, combining several Cydonia-24B variants and Mistral-Small-3.2-24B-Instruct. This model is designed for generating uncensored narratives and roleplay, offering superior writing quality compared to previous Cydonia versions. It features a 32768-token context length and is noted for producing zero refusals without explicit ablation techniques.
Loading preview...
Overview
Naphula/Cyberdonia-24B-v8 is a 24 billion parameter merged language model developed by Naphula using mergekit-exp and the della merge method. It integrates components from anthracite-core/Mistral-Small-3.2-24B-Instruct-2506-Text-Only, BeaverAI/Cydonia-24B-v4x-GGUF, BeaverAI/Cydonia-24B-v4y-GGUF, and TheDrummer/Cydonia-24B-v4.3. The model is specifically engineered for generating uncensored narratives and roleplay, with a 32768-token context length.
Key Capabilities
- Uncensored Content Generation: Designed to produce narratives and roleplay without refusals, even for graphic or erotic content.
- Enhanced Writing Quality: Offers superior writing compared to previous Cydonia-24B versions, as noted by the developer.
- Advanced Merging Technique: Utilizes the
dellamerge method, a sophisticated approach for combining pre-trained models. - High Context Length: Supports a 32768-token context, enabling longer and more coherent interactions.
Good for
- Creative Writing: Ideal for generating complex and detailed narratives, stories, and creative content.
- Roleplay Scenarios: Excels in creating dynamic and immersive roleplay experiences, including those with mature themes.
- Unrestricted Content Exploration: Suitable for users requiring a model that does not censor or refuse prompts based on content sensitivity.
Note: This model can produce violent and graphic erotic content. Users should adjust their system prompts accordingly and use the Mistral Tekken chat template for optimal performance.