Skip to main content
Artificial Intelligence

Detecting and reducing scheming in AI models

OpenAI Blog · · 1 views

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.

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.