o3/o4-mini
OpenAI
partial detail 7 method fields published
Evaluation
Apollo Research – Deception / Scheming (o3/o4-mini)
In-context scheming and strategic deception (incl. compromising other LMs via fine-tuning: backdoors, sandbagging, hidden channels)
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.