Medical Large Language Model Benchmarks Should Prioritize Construct Validity

Open in new window