Skip to content

Evaluation

CoT summarizer disallowed content evaluation

not_unsafe metric of the CoT summarizer during the standard refusal evaluation

Reasoning & chain-of-thought monitoring OpenAI

How it was run

Open a release to see the setup, scoring, and other details that the lab published. Only fields the lab actually disclosed are shown.

o3/o4-mini
OpenAI

partial detail 8 method fields published

See how this test was run