Philipp Emanuel Weidmann @p-e-w CLI
Heretic
Fully automatic censorship removal for transformer language models, without expensive post-training.
- Type
- CLI
- Updated
- 2026-09-17
Install
via Positronick CLI
$positronick cli install heretic --run
Official
$pip install -U heretic-llm
Heretic is a command-line tool that removes refusal/safety alignment from transformer-based language models by combining directional ablation (abliteration) with a TPE parameter optimizer (Optuna). It co-minimizes refusals and KL divergence from the original model so the result stays as capable as possible. Supports most dense models, many multimodal models, several MoE architectures, and some hybrids such as Qwen3.5. Install with pip, then run `heretic <hf-model-id>`. Free software under AGPL-3.0-or-later by Philipp Emanuel Weidmann.