1 paper
Xiaotian Luo, Xun Jiang, Jiangcheng Wu
Interactive medical dialogue benchmarks have shown that LLM diagnostic accuracy degrades significantly when interacting with non-cooperative patients, yet existing approaches eithe…