Knowing Is Not Doing: A Benchmark for Therapeutically Relevant Errors of Large Language Models
Open the PDF