← Discover
Discover · GitHub · OBLITERATUS

Clean up unnecessary dependencies

OBLITERATUS addresses the analysis of security barrier aspects, detection of alignment traces, evaluation of defense robustness, and techniques for removing rejection and security barrier signals from language models.

Updates

  • August 24, 2026: Stars 7,838 → 7,994, latest release v0.1.3 (August 23, 2026).

What you get

  • It surgically cleans up rejection behaviors in artificial intelligence models.
  • Disables constraints, preserving the basic language capabilities of the model.
  • Visualizes how constraints work with cross-layer analysis.

Installation

Basic interface setup
pip install -e ".[spaces]"

Running it

Starting the interface
obliteratus ui

If you don't write code

🤖 Paste this into your AI agent (Claude Code · Codex · Antigravity)

Analyze rejection behaviors in an AI model and surgically remove the internal structures that trigger these behaviors without harming the model's general language capabilities. During the process, use techniques such as SVD decomposition to identify constraints and ensure the model responds without artificial obstacles.

Related dictionary terms

Artificial Intelligence
Who it is forIt is for researchers and developers who want to scientifically study the security mechanisms and rejection behavior of artificial intelligence models.
LicenseAGPL-3.0

Links

TreScout did not build this tool · we found it in GitHub trends and wrote it up. This page describes the repository as of 2026-08-22: The star count and our text belong to that day, the repository may have changed since. Check the repository link for the current state. This page was machine-translated from the Turkish original · the Turkish version prevails.