BPQA (Blood Pressure Question Answering)
收藏资源简介:
BPQA数据集是由纽约大学数据科学中心等机构创建,包含100个经过医学生验证的医学问答对,旨在评估语言模型在理解和使用血压读数方面的能力。数据集覆盖了不同类型的临床问题,包括特殊情境下的异常检测、症状与疾病、干预意见以及医学研究等,每个类别包含25个问答对。该数据集专门设计为依赖血压读数,以评估语言模型在医学问答任务中利用血压信息的能力。
The BPQA dataset was created by the NYU Center for Data Science and other institutions. It includes 100 medical question-answer pairs verified by medical students, and is intended to evaluate the ability of language models to understand and utilize blood pressure readings. The dataset covers diverse types of clinical questions, such as anomaly detection in special scenarios, symptoms and diseases, intervention recommendations, and medical research, with 25 question-answer pairs in each category. Specifically, this dataset is designed to rely on blood pressure readings as core input, aiming to assess language models' capability to leverage blood pressure-related information in medical question-answering tasks.




