中华中医药学刊2025，Vol.43Issue(10)：19-27,后插2,10.DOI:10.13193/j.issn.1673-7717.2025.10.004
面向中医药领域专业能力的生成式大语言模型对比研究
Comparative Study of Generative Large Language Models for Professional Abilities in Traditional Chinese Medicine
摘要
Abstract
Objective To evaluate the knowledge and clinical capabilities of large language models in the field of traditional Chinese medicine.Methods Using literature research and experimental research methods,9 different categories of large language models were selected,a data set of knowledge ability and clinical ability was constructed,and traditional Chinese medicine profes-sional ability prompts were designed.The automatic evaluation method and expert scoring method were used to evaluate the large language models,and the variance was used.Analysis and multiple comparison methods were used to comparatively analyze the medical professional ability level of large language models.Results In the evaluation of knowledge ability,Lingyi Wanwu(78.93),Zhipu Qingyan(77.91)and Tongyi Qianwen(77.22)performed best.There were significant differences in the knowl-edge ability scores of different models,different subjects and different types of models(P＜0.05).In the clinical capability eval-uation,the responses generated by each model had high legibility(81.00),but low accuracy(74.86).Among them,the average score of Wenxin Yiyan was the highest at 85.40,and all five-dimension scores reached excellent levels(≥80).The difference in clinical ability scores of different types of models was significant.Conclusion The general large language model has advantages in generalization,legibility and security.The medical large language model performs well overall in the field of traditional Chinese medicine specialization and traditional Chinese medicine consultation.The application of large language models in the field of tra-ditional Chinese medicine has broad development prospects in the future.关键词
生成式大语言模型/中医药领域/专业能力/对比Key words
generative large language model/the field of traditional Chinese medicine/professional ability/comparison分类
医药卫生引用本文复制引用
张童,王一凡,王若佳,田甜,闫占峰,郭凤英..面向中医药领域专业能力的生成式大语言模型对比研究[J].中华中医药学刊,2025,43(10):19-27,后插2,10.基金项目
国家自然科学基金青年科学基金项目(82204963) （82204963）
教育部中国高校产学研创新基金项目(2021LDA12004) （2021LDA12004）
北京中医药大学教育科学研究项目(XJY22045) （XJY22045）
