LLM-powered medical education dataset

该数据集包含了基于四个标准：专业度、医学相关性、伦理行为和情境干扰，对学生与人工智能患者对话脚本进行模糊标注的对话记录。该数据集主要用于训练和评估作为模糊评委的大型语言模型（LLM-as-a-Fuzzy-Judge），重点是使自动化评估与人工评判的准确性保持一致。该任务旨在自动化评估医学生临床沟通技能。

This dataset comprises dialogue transcripts between medical students and AI patients, with fuzzy annotations generated based on four criteria: professionalism, medical relevance, ethical conduct, and situational distraction. It is primarily intended for training and evaluating Large Language Models acting as Fuzzy Judges (LLM-as-a-Fuzzy-Judge), with the key objective of aligning the accuracy of automated assessments with that of human judgments. This task targets the automated evaluation of clinical communication skills among medical students.