TCM-3CEval
收藏资源简介:
TCM-3CEval是由上海中医药大学、中国科学院上海人工智能实验室和中国中医科学院共同创建的数据集,旨在评估大型语言模型在中医领域的核心知识掌握、经典文本理解和临床决策制定三个维度的表现。该数据集包含450个条目,来源于中医教材、经典文献和临床案例,涵盖了中医基础理论、诊断、中药学、方剂学等多个方面的知识,以及临床诊断和治疗能力。数据集的设计旨在为中医领域的大型语言模型提供科学和系统的评价框架,以促进中医理论与现代技术的深度融合。
TCM-3CEval is a dataset jointly developed by Shanghai University of Traditional Chinese Medicine, Shanghai Artificial Intelligence Laboratory of the Chinese Academy of Sciences, and China Academy of Chinese Medical Sciences. It aims to evaluate the performance of large language models (LLMs) across three core dimensions in the field of traditional Chinese medicine: mastery of core TCM knowledge, comprehension of classic TCM texts, and clinical decision-making. This dataset contains 450 entries sourced from TCM textbooks, classic medical literatures, and clinical cases, covering knowledge in multiple areas including basic TCM theory, diagnostics, Chinese materia medica, formulology, as well as clinical diagnosis and treatment capabilities. The dataset is designed to provide a scientific and systematic evaluation framework for large language models in the TCM domain, so as to promote the in-depth integration of traditional Chinese medicine theories and modern technologies.

- 1TCM-3CEval: A Triaxial Benchmark for Assessing Responses from Large Language Models in Traditional Chinese Medicine上海中医药大学, 中国科学院上海人工智能实验室, 中国中医科学院 · 2025年



