Skip to content

Topic

Frontier Model Safeguards

The refusal policies, classifiers and access restrictions AI developers apply to their most capable models, and the capability thresholds used to justify them.

Current clusters