1 paper · 1 filter
Omid Bazgir, Md Nasir, Jacob Hoffman +6
Synthetic clinical benchmarks for enterprise AI agents can pass existing utility checks and still remain structurally unrealistic, especially in privacy-sensitive healthcare settin…