FlareRebellion/DarkHazard-v1.1-24b
DarkHazard-v1.1-24b is a 24 billion parameter merged language model created by FlareRebellion using the Model Stock method with arcee-ai/Arcee-Blitz as its base. Inspired by Yoesph/Haphazard-v1.1-24b, this model integrates several specialized models to enhance capabilities such as storytelling, roleplay, prompt adherence, and uncensored content generation. With a 32768 token context length, it is designed for diverse creative and conversational applications requiring nuanced responses.
Loading preview...
DarkHazard-v1.1-24b: A Merged Language Model
DarkHazard-v1.1-24b is a 24 billion parameter language model developed by FlareRebellion, built upon the arcee-ai/Arcee-Blitz base model. This model was created using the advanced Model Stock merge method, drawing inspiration from Yoesph/Haphazard-v1.1-24b.
Key Capabilities
This merge integrates several distinct models, each contributing specialized strengths:
- Storytelling and Roleplay: Incorporates
aixonlab/Eurydice-24b-v2to enhance narrative generation and role-playing scenarios. - Prompt Adherence: Leverages
PocketDoc/Dans-PersonalityEngine-V1.2.0-24bfor improved consistency in following user instructions and maintaining specific personas. - Uncensored Content Generation: Includes
ReadyArt/Safeword-Abomination-of-Omega-Darker-Gaslight_The-Final-Forgotten-Transgression-24Bto facilitate less restricted content creation. - General Performance: Benefits from the base model and
Yoesph/Haphazard-v1.1-24bfor overall language understanding and generation.
Good For
- Applications requiring robust story generation and role-playing capabilities.
- Use cases where strict adherence to prompts and defined personalities is crucial.
- Scenarios demanding a model with fewer content restrictions.
- Developers seeking a versatile 24B parameter model with a 32768 token context length for complex conversational and creative tasks.