obliteratus

Remove refusal directions from open-weight LLMs via CLI ablation.

Updated Apr 23, 2026

Search Tags:#llm #svd #weights #guardrails #ablation #mechanistic-interpretability #obliteration

Ready to run

npx skills add https://github.com/Chris-Chai-Minjae/hermes-agent-r1-bridge --skill obliteratus-chris-chai-minjae

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: obliteratus
Source: https://github.com/Chris-Chai-Minjae/hermes-agent-r1-bridge/tree/main/skills/mlops/inference/obliteratus
Command: npx skills add https://github.com/Chris-Chai-Minjae/hermes-agent-r1-bridge --skill obliteratus-chris-chai-minjae

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires obliteratus, torch, transformers, bitsandbytes, accelerate, safetensors, and includes references (resource) components.

What problem does it solve?

Remove refusal behaviors from open-weight LLMs to uncensor or refine their reasoning without retraining.

Core Features & Use Cases

Mechanistic ablation: Remove refusal directions in weights using SVD, LEACE, SAE, and related techniques, enabling targeted removal while preserving core capabilities.
Safety-first workflow: Maintains reasoning quality and controllable behavior through telemetry-driven recommendations and model presets.
Use case: Researchers can experiment with uncensored variants of large models to study guardrail mechanisms and evaluate alignment trade-offs.

Quick Start

Invoke the CLI to run the abliteration workflow on your target model.

Frequently Asked Questions about obliteratus

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I remove refusal behaviors from open-weight LLMs without retraining?▼

To remove refusal behaviors from open-weight LLMs without retraining, you can apply mechanistic ablation techniques. This Skill uses SVD, LEACE, and SAE methods to modify weights directly, targeting refusal directions while preserving core model capabilities.

What is mechanistic interpretability for uncensoring an LLM?▼

Mechanistic interpretability for uncensoring an LLM involves identifying and ablating refusal directions within the model weights. This Skill applies SVD and related techniques to remove guardrails safely, producing an abliterated model with verification metrics.

Do I need specific ML libraries to run abliteration on my model?▼

Yes, you need specific ML libraries to run abliteration on your model. This workflow requires torch, transformers, bitsandbytes, accelerate, and safetensors to perform the weight modifications and load open-weight models properly.

Can I use SVD to modify model weights for alignment research?▼

Yes, you can use SVD to modify model weights for alignment research. Researchers can experiment with uncensored variants of large models to study guardrail mechanisms and evaluate alignment trade-offs using the provided telemetry-driven recommendations and model presets.

Does ablating refusal directions affect the reasoning quality of the model?▼

Ablating refusal directions is designed to maintain reasoning quality while removing guardrails. The safety-first workflow uses targeted weight modification to ensure controllable behavior, preserving core capabilities through verified metrics after the abliteration process.