Open Issues Need Help
View All on GitHub [Research-grade compute contribution] Full Security-First V5 training + reproducible matched evaluation about 3 hours ago
help wanted
A counter-espionage RL environment for scalable AI oversight. Train LLM agents to detect deception, trace leaks, and neutralize sleeper agents in a 160-turn corporate simulation. Built for the Meta PyTorch OpenEnv Hackathon.
Python
#ai-safety#deception-detection#hackathon#llm-agents#multi-agent#openenv#pytorch#reinforcement-learning