Google launches a pilot of double-blind AI evaluations, keeping external evaluations in a cryptographic "box" to stop benchmark contamination and protect IP (Google DeepMind)

1 month ago 21
Add to circle

Google DeepMind:
Google launches a pilot of double-blind AI evaluations, keeping external evaluations in a cryptographic “box” to stop benchmark contamination and protect IP  —  Building trust in proprietary model benchmarks using cryptographically secure environments  —  Imagine a student is set to take a high-stakes exam.

Read Entire Article