ChatCoach

ChatCoach数据集是由新加坡国立大学创建的，旨在通过模拟医疗咨询环境，帮助医学学习者提升沟通技巧。该数据集包含3500个对话，总计13666条语句，分为训练、验证和测试集。数据集的创建过程采用了多代理数据生成框架，利用外部资源如医疗对话数据库和医疗知识数据库来合成训练数据。ChatCoach数据集特别适用于评估大型语言模型在医疗沟通辅导任务中的有效性，旨在解决医疗专业人员沟通技能培训的需求。

The ChatCoach dataset was developed by the National University of Singapore to help medical learners enhance their communication skills by simulating medical consultation environments. It contains 3500 conversations totaling 13666 utterances, and is split into training, validation, and test subsets. The dataset was created using a multi-agent data generation framework, leveraging external resources such as medical dialogue databases and medical knowledge bases to synthesize training data. The ChatCoach dataset is particularly suitable for evaluating the effectiveness of Large Language Models (LLMs) in medical communication coaching tasks, and aims to address the demand for communication skill training among medical professionals.