invisietch/L3.3-Ignition-v0.1-70B
TEXT GENERATIONConcurrency Cost:4Model Size:70BQuant:FP8Ctx Length:32kPublished:Aug 27, 2025License:llama3.3Architecture:Transformer0.0K Cold

invisietch/L3.3-Ignition-v0.1-70B is a 70 billion parameter Llama 3.3-based model merge developed by invisietch, specifically designed and optimized for creative roleplay and fiction writing. It underwent a multi-stage merge process to enhance creative writing capabilities, minimize 'slop,' and improve coherence compared to its constituent models. The model demonstrates a preference for detailed character cards and is sensitive to specific system prompting, with coherent inference tested up to approximately 64k context.

Loading preview...