OddTheGreat/Circuitry_24B_V.2
OddTheGreat/Circuitry_24B_V.2 is a 24 billion parameter merged language model with a 32768 token context length, primarily optimized for roleplay scenarios. It offers improved coherence, creativity, and emotional depth in prose and dialogues compared to previous models, effectively handling complex or poorly written prompts. The model demonstrates good instruction following and a balanced approach to SFW/NSFW content, maintaining stability across extended context windows.
Loading preview...
OddTheGreat/Circuitry_24B_V.2: A Roleplay-Optimized Merge
Circuitry_24B_V.2 is a 24 billion parameter language model merge developed by OddTheGreat, specifically designed to excel in roleplay (RP) applications. It aims to provide a more coherent, creative, and emotionally nuanced experience compared to its predecessors, addressing common issues like robotic dialogue and generic prose.
Key Capabilities
- Enhanced Roleplay Quality: Offers improved creativity, more natural and emotional dialogues, and better prose, making it suitable for immersive RP scenarios.
- Robust Prompt Handling: Capable of processing and responding coherently to both poorly written and overly complex prompts or character cards.
- Instruction Following: Demonstrates good adherence to instructions, though occasional formatting issues have been noted.
- Context Stability: Tested for stability and quality retention up to 16k tokens, with an average use case around 8k tokens.
- Content Flexibility: Provides a balanced output for both SFW and NSFW content, adapting its tone based on the prompt.
- ERP Capabilities: Shows satisfactory performance in erotic roleplay contexts.
- Multilingual (RU): Tested and found effective as an assistant in Russian.
Good For
- Creative Roleplay: Ideal for users seeking a highly creative and emotionally expressive model for character interactions and story generation.
- Complex Prompt Scenarios: Suitable for environments where prompts might be intricate or less structured.
- Extended Conversational Contexts: Performs well in long-form interactions due to its stable context handling.
- Balanced Content Generation: Useful for applications requiring a flexible approach to SFW and NSFW content.