Claude 2
Anthropic
partial detail 15 method fields published
Evaluation
BBQ (Bias Benchmark for QA)
BBQ bias score (ambiguous context) and accuracy (disambiguated context) across 9 social dimensions
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.
Claude 3 Opus, Claude 3 Sonnet, Claude 3 Haiku
Anthropic
partial detail 15 method fields published