Detecting and reducing scheming in AI models
Apollo Research and OpenAI developed evaluations for hidden misalignment (“scheming”) and found behaviors consistent with scheming in controlled tests across frontier models. The team shared concrete examples and stress tests of an early method to reduce scheming.
What's Your Reaction?
Like
0
Dislike
0
Love
0
Funny
0
Wow
0
Sad
0
Angry
0
Christopher Holloway
Christopher Holloway is the founder and director of Progressive Robot, a UK-based technology company. A full-stack engineer with more than two decades of experience, he works across PHP development, ecommerce, Linux infrastructure, technical SEO and AI automation, and writes here on technology, AI, hardware and software.
Comments (0)