Congbo Ma

General LLMs outperform medical models in new multilingual benchmark

Researchers have developed MedErrBench, a novel multilingual benchmark for evaluating large language models' ability to detect, locate, and correct errors in medical texts. The study, which included English, Chinese, an…