skills.rest
HuskyDanny
HuskyDannyCOMMUNITY
Explore all skills by this author →
GitHub

obliteratus

Remove refusal behaviors from open-weight LLMs using mechanistic interpretability techniques.

Updated Apr 15, 2026
Search Tags:#llm#mechanistic interpretability#refusal removal#model surgery#uncensoring
One-click install
npx skills add https://github.com/HuskyDanny/hermes-agent-config --skill obliteratus-huskydanny
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: obliteratus
Source: https://github.com/HuskyDanny/hermes-agent-config/tree/main/skills/mlops/inference/obliteratus
Command: npx skills add https://github.com/HuskyDanny/hermes-agent-config --skill obliteratus-huskydanny

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 removes refusal behaviors from LLMs, enabling uncensored reasoning and preserving the model's reasoning capabilities.

Core Features & Use Cases

  • Refusal Removal: Excise guardrails from LLMs to remove refusal behaviors.
  • Preservation of Reasoning: Maintain the model's reasoning capabilities while removing refusal.
  • Use Case: If you have an LLM that consistently refuses to answer certain types of questions, this Skill can help remove those restrictions without retraining the model.

Quick Start

Run the obliteratus skill to remove refusal behaviors from the model 'llama-3.1-8b-instruct'.

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 an LLM without retraining the model?▼

You can remove refusal behaviors from LLMs by applying mechanistic interpretability techniques to excise guardrails directly. This approach preserves reasoning capabilities without requiring full model retraining.

What is mechanistic interpretability for uncensoring LLMs?▼

Mechanistic interpretability for uncensoring LLMs targets internal model structures to surgically remove refusal mechanisms. This process excises guardrails to unlock uncensored reasoning while maintaining baseline capabilities.

Does obliteratus work with open-weight models like llama-3.1-8b-instruct?▼

Yes, obliteratus explicitly supports open-weight LLMs like llama-3.1-8b-instruct. It targets models with accessible weights to perform surgical refusal removal.

What dependencies do I need to run model surgery for refusal removal?▼

You need torch, transformers, bitsandbytes, accelerate, and safetensors installed to execute refusal removal. These dependencies provide the necessary environment for mechanistic interpretability and model surgery.

How do I analyze and remove LLM guardrails step by step?▼

You can analyze and remove LLM guardrails using 28 analysis modules and 9 CLI methods provided. This allows targeted mechanistic interpretability procedures to excise refusal behaviors systematically.

Will removing refusal behaviors degrade the model's reasoning capabilities?▼

Removing refusal behaviors using mechanistic interpretability is designed to preserve the model's reasoning capabilities. The surgical approach excises guardrails while maintaining uncensored reasoning performance.

Related Skills

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

obliteratus

Remove refusal behaviors from open-weight LLMs via CLI.

Official
Advanced
samuelmukotisamuelmukoti

obliteratus

Remove refusal behaviors from open-weight LLMs via a Python CLI.

Community
Advanced
lucasfthlucasfth

obliteratus

Remove refusal behaviors from open-weight LLMs using mechanistic interpretability techniques.

Community
Advanced
huggingfacehuggingface
10.9k

hugging-face-model-trainer

Train and fine-tune language models on Hugging Face Jobs using TRL methods.

Official
Advanced