How we detect frontier model misalignment
AI Summary
OpenAI uses layered monitoring to detect frontier model misalignment and unauthorized agent behavior so responders can intervene before incidents escalate.
Discussion
Sign in to join the discussion.
Loading comments…