skills.rest
zhouboyu-xreal
zhouboyu-xrealCOMMUNITY
Explore all skills by this author →
GitHub

obliteratus

Removes LLM refusal behaviors by excising refusal directions from model weights using obliteratus, torch, and SVD-based methods.

Updated Jun 9, 2026
Search Tags:#LLM#AI safety#mechanistic interpretability#refusal removal#model surgery
One-click install
npx skills add https://github.com/zhouboyu-xreal/Hermes-Memory --skill obliteratus-zhouboyu-xreal
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: obliteratus
Source: https://github.com/zhouboyu-xreal/Hermes-Memory/tree/main/skills/mlops/inference/obliteratus
Command: npx skills add https://github.com/zhouboyu-xreal/Hermes-Memory --skill obliteratus-zhouboyu-xreal

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

This Skill allows users to remove refusal behaviors from LLMs without retraining, enabling more open and responsive language models.

Core Features & Use Cases

  • Refusal Removal: Eliminate refusal behaviors (guardrails) from LLMs without retraining or fine-tuning.
  • Mechanistic Interpretability: Utilizes techniques like diff-in-means, SVD, and whitened SVD to identify and remove refusal directions from model weights.
  • Use Case: For users who want to create uncensored versions of LLMs or analyze the refusal mechanisms of models.

Quick Start

Run the obliteratus command with the model name and method you want to use.

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 LLMs without retraining?▼

You can remove refusal behaviors from LLMs by identifying and excising refusal directions directly from model weights. This approach uses mechanistic interpretability techniques to alter the model without requiring fine-tuning or retraining.

What techniques are used to identify refusal directions in LLM weights?▼

Refusal directions in LLM weights are identified using mechanistic interpretability techniques such as diff-in-means, SVD, and whitened SVD. These methods isolate the specific weight components responsible for guardrail behaviors.

Can I use this model surgery approach on any open-weight LLM?▼

Yes, this model surgery approach applies to open-weight LLMs. You need to ensure your environment has the required dependencies, including torch, transformers, bitsandbytes, accelerate, and safetensors.

Does eliminating refusal directions require fine-tuning the model?▼

No, eliminating refusal directions does not require fine-tuning. The process performs model surgery directly on the existing weights to uncensor the model and bypass safety guardrails without retraining.

What dependencies do I need to run the obliteratus refusal removal script?▼

To run the obliteratus refusal removal script, you need torch, transformers, bitsandbytes, accelerate, and safetensors installed. These libraries facilitate the weight manipulation and model loading processes.

Related Skills

Explore More Software Engineering Skills→
humanerd-drewhumanerd-drew
2

obliteratus

Remove refusal behaviors from open-weight LLMs via CLI.

Official
Advanced
Simon-Copilot-StudioSimon-Copilot-Studio

obliteratus

Remove refusal directions from open-weight LLM weights using mechanistic interpretability techniques.

Community
Advanced
lxh755818-botlxh755818-bot

obliteratus

Remove refusal behaviors from open-weight LLMs using diff-in-means and SVD.

Community
Advanced
DarkArty07DarkArty07
3

obliteratus

Remove refusal guardrails from open-weight LLMs via weight projection.

Community
Advanced