PSEBench evaluates LLMs for patient safety triage

Arxiv CS.AI (top papers)

PSEBench is a benchmark designed to evaluate LLMs on patient safety event triage, a high-stakes clinical task. It tests whether LLMs can correctly determine reportability under jurisdiction-specific policies.

benchmarks

Arxiv CS.AI (top papers)·Keqi HanRyan YoungAnnabel StraussLindsey HughesKatharine M. NesbittNicole SchuelerChe NguforCarl YangYuan XueZhijun Yin

6 days ago