Announcement
BehaviorBench evaluates foundation models across 20 scenarios and four capabilities
The introductory post asks how well frontier models understand human behavior and links to a leaderboard and paper.
TLDR
A post introducing BehaviorBench says it evaluates foundation models across 20 scenarios and four capabilities. It frames the benchmark around how well frontier models understand human behavior and links to a leaderboard and paper.
Combined views
1.7K
2 Sources, first seen ago
