facebookresearch/advgame

Code for paper "MAGIC: A Co-Evolving Attacker-Defender Adversarial Game for Robust LLM safety"

[NeurIPS 2024] Official implementation for "AgentPoison: Red-teaming LLM Agents via Memory or Knowledge Base Backdoor Poisoning"

Toolbox to run adversarial attacks against LLM.

An autonomous red-teaming engine for LLMs. RedThread manages the full security lifecycle: generating adversarial attacks, executing precision evaluations, and synthesizing validated guardrails for safe self-improvement.

An easy-to-use Python framework to generate adversarial jailbreak prompts.

Code Repository for: AIRTBench: Measuring Autonomous AI Red Teaming Capabilities in Language Models

AutoDefense: Multi-Agent LLM Defense against Jailbreak Attacks

A trivial programmatic Llama 3 jailbreak. Sorry Zuck!

Reproduction code for "A Single Neuron Is Sufficient to Bypass Safety Alignment in Large Language Models"