MyoCardBench: A Real-World Benchmark for Evaluating LLMs in Cardiovascular Care
Researchers created MyoCardBench, a real-world benchmark using 2,263 items from 13 datasets derived from de-identified cardiovascular records, to evaluate how well large language models perform in clinically authentic, longitudinal, and multimodal cardiovascular care scenarios.
