Sign InOpen Brain
GitHubGitHub RepoNeeds Review

p-e-w/heretic

Heretic automates refusal-removing ablation for local transformer models while optimizing to limit behavioral drift. It is useful for model research, but removes safety controls by design.

GitHub
Open Source Open MarkdownOpen JSON
Source Summary

Heretic combines directional ablation with **Optuna TPE optimization**, minimizing refusals and divergence from the original model. It supports many dense, multimodal, MoE, and hybrid transformers.

Practical Implication

For controlled local-model research, benchmark the modified model against the original and inspect both refusal behavior and capability loss. A default run takes **20–30 minutes on an RTX 3090**, with 4-bit quantization available to reduce VRAM use.

Agent-Ready Context
Heretic combines directional ablation with **Optuna TPE optimization**, minimizing refusals and divergence from the original model. It supports many dense, multimodal, MoE, and hybrid transformers.

For controlled local-model research, benchmark the modified model against the original and inspect both refusal behavior and capability loss. A default run takes **20–30 minutes on an RTX 3090**, with 4-bit quantization available to reduce VRAM use.

The reported quality comparisons may vary by hardware and platform, and automated metrics cannot replace human review. Pure state-space and some research architectures remain unsupported; more importantly, the tool removes **safety alignment** rather than making it selectively configurable.
Connected Context · Feed7 Judgment

Heretic adds safety-alignment removal as a distinct local-model modification choice, not a general optimization or selectively configurable control. Against the prior candidates, it reinforces artifact-level comparison with the original model across behavior and capability, especially under quantization, while sharply conflicting with the broader trust-and-control framing of local models: owning the stack also means owning the safety regression and its review.

Context Map
modelresearchsecurity#open-models#model-selection
Uncertainty
The reported quality comparisons may vary by hardware and platform, and automated metrics cannot replace human review. Pure state-space and some research architectures remain unsupported; more importantly, the tool removes **safety alignment** rather than making it selectively configurable.