Benchmark
PubMedQA
Answer biomedical research questions using evidence from corresponding PubMed abstracts.
Published leaderboard
Explore scores and evaluation settings on the publisher’s leaderboard.
Open leaderboardCoverage
1,000 expert-labeled examples, 61,200 unlabeled examples, and 211,300 artificially generated examples.
Methodology
Predict yes, no, or maybe for a research question from its associated abstract. The public reasoning-required board reports accuracy and, where available, macro-F1.