resource · 2025
GPT-4 作答台湾中医师执照考试
JMIR Medical Education
GPT-4 作答台湾中医师执照考试 is a resource in the TCM LLM catalog (Awesome-TCM-LLM), dated 2025. It is maintained alongside the GitHub list of Traditional Chinese Medicine large language models.
GPT-4 answered 480 items from the 2022 Taiwan TCM licensing examination and was compared with two licensed TCM doctors, using direct answers and explain-then-answer. Overall accuracy was 43.9% for the model, versus 70% and 78.4% for the two doctors. About 55.3% of the errors came from misunderstanding TCM concepts. This compares exam recall with licensed practitioners; it does not separately score syndrome judgment, prescription rationality, and safety checks.