bilawalriaz/MedAtoms-Combined

MedAtoms是一个开放的原子医学事实数据集，旨在构建可靠的医疗保健AI系统。这是完整的数据集，结合了来自精选医学文献和维基百科的高置信度原子事实，提供了最全面的覆盖范围。关键特点包括：仅包含高置信度内容、最大覆盖范围、去重处理、灵活性（可通过source_type字段按来源筛选）以及生产就绪（包含训练/验证/测试分割和每个事实的唯一ID）。数据集包含476,684个事实，覆盖35个医学领域，平均每个事实长度为18个单词。

MedAtoms is an open dataset of atomic medical facts designed for building reliable healthcare AI systems. This is the complete dataset combining high-confidence atomic facts from both curated medical literature and Wikipedia, providing the most comprehensive coverage available. Key features include: High confidence only, Maximum coverage, Deduplicated, Flexible (use `source_type` field to filter by origin), and Production ready (train/val/test splits with unique IDs for every fact). The dataset contains 476,684 facts covering 35 medical areas with an average fact length of 18 words.