AI research and discussion on LessWrong. www.lesswrong.com
Proposes a roadmap for safety cases built on chain-of-thought monitoring, covering dangerous-capability evals without CoT, encoded reasoning threats, and methods to preserve CoT monitorability.
Sketches operational interventions, such as model-whisperer roles, rephrasing layers, interaction limits and rollback, to contain mildly superhuman AI persuasion that could enable self-exfiltration, and proposes LLM-based testbeds.
Argues AI companies should log, analyze and real-time monitor internal use of frontier models for scheming and misbehaviour, since internal deployments can threaten security even without external harm.