First double-blind eval of proprietary model via secure enclave

Miles_Brundage · x · 2026-08-28

AVERI, Google DeepMind, OpenMined, and MLCommons achieved the first double-blind evaluation of a proprietary model (Gemini 2.5 Flash-Lite). Conducted inside a secure enclave, the process ensures the lab (GDM) cannot see eval prompts and the evaluator cannot see weights. Cryptographic attestation verifies the correct model and evals were used.

Related event: Google DeepMind pilots first double-blind evaluation of a proprietary frontier model(13 posts)→

Original post →

More from Safety

Safety channel →