stanfordmimi/MedVAL-Bench
收藏Hugging Face2025-09-04 更新2025-08-09 收录
下载链接：
https://hf-mirror.com/datasets/stanfordmimi/MedVAL-Bench
下载链接
链接失效反馈官方服务：
资源简介：
MedVAL-Bench是一个医疗文本评估基准数据集，用于微调和评估语言模型评估AI生成的医疗文本输出的能力。该数据集包括自我监督的训练集和包含医生评估的测试集，支持五种医疗文本生成任务，并提供医生定义的风险等级和错误分类法。
MedVAL-Bench is a medical text assessment benchmark dataset for fine-tuning and evaluating the ability of language models to assess AI-generated medical text outputs. The dataset includes a self-supervised training set and a test set with physician assessments, supporting five medical text generation tasks, and providing physician-defined risk levels and error taxonomies.
提供机构：
stanfordmimi
