mteb/MIRACLRetrieval_hi_top_250_only_w_correct

该数据集包含了三个不同的配置：文本数据集（corpus）、查询和文档匹配数据集（default）以及查询数据集（queries）。文本数据集提供了标题和文本内容，查询和文档匹配数据集包含查询ID、文档ID和匹配得分，查询数据集则只包含文本内容。每个配置都有相应的训练集划分，并提供了相应的数据大小和示例数量信息。

The dataset consists of three different configurations: a text dataset (corpus), a query-document matching dataset (default), and a query dataset (queries). The text dataset provides title and text content, the query-document matching dataset includes query ID, document ID, and matching scores, while the query dataset contains only text content. Each configuration has its respective training set split, along with information on data size and the number of examples.