ReaDiagnosis
收藏资源简介:
ReaDiagnosis数据集由华东理工大学和上海交通大学医学院瑞金医院共同创建,包含2225条从中国医疗网站收集的高质量电子病历记录。该数据集覆盖12个科室,每条记录包含五个与诊断相关的问题,旨在模拟真实临床环境中的多步骤诊断过程。数据集的创建过程包括数据收集、预处理和多轮注释,确保了数据的可靠性和真实性。ReaDiagnosis主要用于评估和提升大型语言模型在复杂临床诊断任务中的表现。
The ReaDiagnosis dataset was jointly developed by East China University of Science and Technology and Ruijin Hospital Affiliated to Shanghai Jiao Tong University School of Medicine. It encompasses 2225 high-quality electronic medical record (EMR) entries collected from Chinese medical websites, covering 12 clinical departments. Each entry features five diagnosis-associated questions, designed to simulate the multi-step diagnostic workflow in real-world clinical scenarios. The dataset's development pipeline includes data collection, preprocessing and multi-round annotation, which ensures the reliability and authenticity of the data. ReaDiagnosis is primarily used to evaluate and enhance the performance of large language models (LLMs) in complex clinical diagnostic tasks.




