OccultAI/MN-Nazgul-12B-v1
OccultAI/MN-Nazgul-12B-v1 is a 12 billion parameter language model merged using the 'della' method, based on Retreatcost/Mistral-Nemo-Base-2407-ChatML. This model is designed to be uncensored and refusal-free, specifically optimized for generating narratives and roleplay content, including potentially violent and graphic erotic material. It leverages a blend of several 12B models to achieve its specialized output, with a context length of 32768 tokens.
Loading preview...
MN Nazgul 12B v1: Uncensored Narrative and Roleplay Model
OccultAI/MN-Nazgul-12B-v1 is a 12 billion parameter language model created through a sophisticated merge of multiple pre-trained models using the della merge method. Built upon Retreatcost/Mistral-Nemo-Base-2407-ChatML, this model is engineered to be fully uncensored and refusal-free, making it suitable for generating a wide range of narrative and roleplay content.
Key Characteristics
- Uncensored Output: Designed to produce content without typical safety filters, including potentially violent and graphic erotic narratives.
- Merge Architecture: Combines the strengths of several 12B models, including
DarkArtsForge/MN-Raven-12B-v1,IggyLux/MN-VelvetCafe-RP-12B-V2,OccultAI/MN-Morpheus-12B-v1,shrugging-shoulders/Amberlight-Lux-12B, andWokeAI/Tankie-DPE-12B-SFT-v2. - ChatML Template: Utilizes the ChatML chat template for optimal interaction.
- Context Length: Supports a context window of 32768 tokens.
Considerations for Use
While designed for uncensored output, users should be aware that the model may struggle with longer context and requires careful system prompt adjustment to manage content generation. Its primary differentiation lies in its ability to generate explicit and unrestricted narrative and roleplay scenarios.