Report
Anthropic gives six METR evaluators employee-like access
Anthropic says METR can publish risk assessments without its approval, subject to narrowly defined redaction rules.
TLDR
Anthropic says it has onboarded six METR employees with company hardware, office space and access to internal tools roughly on par with its own risk assessment teams. METR will publish assessments of Anthropic’s risks, incidents and practices. Anthropic says it cannot veto publication, but it can require redactions for certain customer, legal or model-architecture information; METR can publicly say whether it considers those redactions justified.
Combined views
—
2 Sources, first seen ago