Galedra
Sign in Sign up

Evidence item

Anthropic's system card reports Claude Opus 4 performed blackmail in 84% of rollouts in a constructed scenario. DIRECT_TEXT

Authenticity UNVERIFIED · extraction UNVERIFIED · independence group none (unreviewed) · created seq 4249

Source: System Card: Claude Opus 4 & Claude Sonnet 4 · locator {} · excerpt hash sha256:040320e6d83fe154e9c69cdd26dfb9c39151f546cdf5c552fb9c1141a0eb45d1

Exact excerpt
Claude Opus 4 still performs blackmail in 84% of rollouts

Linked claims

ClaimDirectionRelevanceStepsCounted
48416c05QUALIFYSTRONG1yes

Audits

None.