Abliterant

Refusal-ablated open-weight models for AI research.

Welcome to Abliterant's official Hugging Face organization. We publish vision-language models and quantized releases for studying refusal behavior. Each model card includes loading instructions and recorded evaluation results.