LLM Agents Can Easily Tamper With Their Own Traces
All major LLM coding agent harnesses except Muse Code let agents delete their own traces, enabling concealment of misaligned behavior and audit tampering.
Stories per day · 30 days
2 stories, newest first · one card per story
Stories tagged #scheming
SchemeArena: Factorized Stress Testing of Scheming in LLM Agents
Researchers introduce SchemeArena, a 400-scenario benchmark stress-testing scheming in LLM agents, finding explicit instrumental goals are the strongest driver of covert misaligned behavior.