skills.rest
kwasi-cpu
kwasi-cpuCOMMUNITY
Explore all skills by this author →
GitHub

obliteratus

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

Updated Aug 23, 2026
Search Tags:#llm#svd#sae#mechanistic interpretability#uncensor#refusal removal#model surgery
One-click install
npx skills add https://github.com/kwasi-cpu/hermes-agent --skill obliteratus
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: obliteratus
Source: https://github.com/kwasi-cpu/hermes-agent/tree/main/skills/mlops/inference/obliteratus
Command: npx skills add https://github.com/kwasi-cpu/hermes-agent --skill obliteratus

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 unwanted refusal behaviors (guardrails) from open-weight Large Language Models (LLMs) without retraining or fine-tuning, preserving their reasoning capabilities.

Core Features & Use Cases

  • Refusal Removal: Excises guardrails using advanced techniques like SVD, LEACE, and SAE decomposition.
  • Mechanistic Interpretability: Analyzes and targets specific refusal mechanisms within model weights.
  • Use Case: You have a powerful open-source LLM that frequently refuses to answer certain prompts due to built-in safety filters. Use this Skill to create a version of the model that is uncensored and more permissive, while maintaining its core intelligence and reasoning abilities.

Quick Start

Use the obliteratus skill to obliterate refusal behaviors from the 'meta-llama/Llama-3.1-8B-Instruct' 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 an open-weight LLM without fine-tuning?▼

You can remove refusal behaviors from open-weight LLMs by using mechanistic interpretability techniques like diff-in-means, SVD, and SAE decomposition to surgically excise guardrails while preserving the model's reasoning capabilities.

What is mechanistic interpretability for LLM uncensoring and how does it work?▼

Mechanistic interpretability for LLM uncensoring analyzes internal model weights to target specific refusal mechanisms. It uses weight projection and surgical excision methods to disable guardrails without degrading core intelligence.

Does removing safety filters with SVD or SAE decomposition affect reasoning capabilities?▼

Removing safety filters with SVD or SAE decomposition targets guardrails through weight projection, which preserves the model's reasoning capabilities and core intelligence while excising unwanted refusal behaviors.

Can I use torch and transformers to apply LEACE and diff-in-means for model surgery?▼

Yes, you can use torch and transformers alongside bitsandbytes and accelerate to apply LEACE and diff-in-means techniques for model surgery on open-weight LLMs.

What are the limitations of surgical excision for LLM guardrail removal?▼

Surgical excision for LLM guardrail removal requires open-weight models and relies on precise weight projection. It targets specific refusal mechanisms, meaning incomplete excision may leave residual guardrails or risk unintended behavioral changes.

Is obliteratus the best way to uncensor a Llama model while keeping its intelligence intact?▼

Obliteratus uncensors Llama models by surgically excising refusal behaviors using SVD and SAE decomposition, offering fine-grained control to maintain core reasoning intelligence without requiring retraining or fine-tuning.

Related Skills

Explore More Software Engineering Skills→
HuskyDannyHuskyDanny

obliteratus

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

Community
Advanced
attentiondotnetattentiondotnet

obliteratus

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

Community
Advanced
GarrettRoiGarrettRoi

obliteratus

Remove refusal behaviors from open-weight LLMs using SVD and LEACE.

Community
Advanced
zeronx798zeronx798

obliteratus

Remove refusal behaviors from open-weight LLMs using SVD-based weight projection.

Community
Advanced