ChaoticNeutrals/Captain_Eris_Noctis-12B-v0.420
Captain_Eris_Noctis-12B-v0.420 is a 12 billion parameter language model developed by ChaoticNeutrals, built using a slerp merge of Nitral-AI/Nera_Noctis-12B-v0.420 and Nitral-AI/Captain-Eris-Diogenes_Twilight-V0.420-12B. This model is designed for conversational AI, supporting a 32768 token context length and utilizing the ChatML prompt format. It is particularly optimized for interactive text generation and roleplay scenarios, with specific presets available for platforms like SillyTavern.
Loading preview...
Captain_Eris_Noctis-12B-v0.420 Overview
Captain_Eris_Noctis-12B-v0.420 is a 12 billion parameter language model developed by ChaoticNeutrals, specifically engineered for interactive and conversational AI applications. It leverages a slerp merge technique, combining the strengths of Nitral-AI/Nera_Noctis-12B-v0.420 and Nitral-AI/Captain-Eris-Diogenes_Twilight-V0.420-12B, with a specific t parameter configuration for self-attention and MLP layers, and a global t value of 0.420. The model processes inputs in bfloat16 data type.
Key Capabilities
- ChatML Prompt Format: Designed to work seamlessly with the ChatML format, facilitating clear system, user, and assistant turns in conversations.
- Extended Context Window: Supports a substantial context length of 32768 tokens, allowing for more coherent and extended dialogues.
- Optimized for Interactive Text Generation: The model's architecture and merging strategy suggest an optimization for generating dynamic and engaging text, particularly suitable for roleplay and narrative applications.
Use Cases
- Conversational Agents: Ideal for building chatbots and virtual assistants that require nuanced and context-aware responses.
- Roleplay and Creative Writing: Specifically tailored with available presets for platforms like SillyTavern, indicating strong performance in generating character-driven narratives and interactive roleplay scenarios.
- Customizable Presets: Comes with pre-configured instruct/context import and Textgen presets, simplifying deployment and fine-tuning for specific interactive text generation tasks.