Claude Opus 4, Claude Sonnet 4
Anthropic
partial detail 17 method fields published
Evaluation
Task preferences experiment (Claude Opus 4)
Elo ratings of tasks from pairwise selections over 75 rounds; strongest preference against harmful tasks (87.2% of harmful tasks negatively rated vs 7.9% of positive-impact tasks)
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.