deu05232/promptriever-RQ1-same_version

该数据集是一个用于信息检索或问答任务的结构化数据集，包含查询ID、查询文本、正例文档列表（含文档ID、解释、FollowIR评分、联合ID、文本和标题）、负例文档列表（含文档ID、文本和标题）、仅指令字段、仅查询字段、是否包含指令标志以及新增负例列表（结构与正例类似）。数据集旨在支持模型训练，以区分相关和不相关文档，并可能用于解释生成或评分预测。数据量较大，训练集包含约423,706个示例。This dataset is a structured dataset for information retrieval or question answering tasks. It includes query ID, query text, positive document list (containing document ID, explanation, FollowIR score, joint ID, text and title), negative document list (containing document ID, text and title), instruction-only field, query-only field, instruction inclusion flag, and additional negative example list which has a structure similar to that of the positive examples. The dataset aims to support model training for distinguishing relevant and irrelevant documents, and can also be potentially used for explanation generation or score prediction. It has a large scale, with approximately 423,706 examples in the training set.