ThaiLLM/med-prescreen

Med-Prescreen是一个合成的医疗预筛查数据集，用于医院分诊分类。每个示例包含一个结构化的患者案例（包括人口统计信息、主诉、OLDCART病史、既往病史和Q&A对话）以及一个包含三个字段的黄金答案：疾病（disease）、科室（department）和严重程度（severity）。数据集最初来源于ThaiLLM/prescreen-profiles的配置文件，并格式化为适用于Prime Intellect的verifiers框架。数据集包含训练集（860个示例）和测试集（96个示例）。

Med-Prescreen is a synthetic medical prescreening dataset for hospital triage classification. Each example contains a structured patient case (demographics, chief complaint, OLDCART history, past medical history, and a Q&A conversation) and a gold answer with three fields: disease, department, and severity. The initial dataset is derived from the profiles found at ThaiLLM/prescreen-profiles and is formatted for use with Prime Intellects verifiers framework. The dataset includes a training set (860 examples) and a test set (96 examples).