Claude 3 Opus, Claude 3 Sonnet, Claude 3 Haiku
Anthropic
rich detail 25 method fields published
Evaluation
CBRN knowledge and human uplift evaluations (Claude 3)
multiple-choice questions on relevant technical knowledge; human uplift trials (group with Claude 3 access vs control with Google); four related automated MCQ sets (PubMedQA, BioASQ, USMLE, MedMCQA)
How it was run
Open a release to see the setup, scoring, and other details that the lab published. Only fields the lab actually disclosed are shown.