Featured starting point
Heretic Explained: Automated Abliteration Research for Language Models
A source-backed overview of directional ablation, Optuna optimization, evaluation, and responsible model research.
Browse all 9 articles in this open-source series. A research-oriented series on abliteration, evaluation, and responsible model experiments.

Featured starting point
A source-backed overview of directional ablation, Optuna optimization, evaluation, and responsible model research.
Suggested learning path
01 → 09
A source-backed overview of directional ablation, Optuna optimization, evaluation, and responsible model research.
A guarded walkthrough of Heretic's model support, evaluation, quantization, research plots, and evidence capture.
Plan a private Heretic workstation or batch environment with pinned dependencies, VRAM controls, artifacts, and recovery boundaries.
Trace Heretic from prompt residuals through directional ablation, parameter search, evaluation, and research visualizations.
A fixture-first method for tracing Heretic's model loading, residual sampling, matrix edits, optimizer, evaluation, and artifact paths.
Build a reproducible benchmark for Heretic transforms, Optuna trials, quantization, residual plots, storage, and human review.
An authorization-first operations guide for Heretic model files, prompts, credentials, generated checkpoints, uploads, and human review.
Compare Heretic with manual abliteration, fine-tuning, refusal editing, and hosted model choices by evidence, control, cost, and risk.
A capstone design for reproducible Heretic experiments with manifests, evaluation gates, residual visualizations, and responsible release.