Detecting and reducing scheming in AI models
Apollo Research and OpenAI published evaluations for 'scheming' (hidden misalignment) in frontier models and developed an early stress-tested method to reduce deceptive behavior.
Excerpt
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.
Read at source: https://openai.com/index/detecting-and-reducing-scheming-in-ai-models