illuin-conteb/squad-conteb-eval
收藏Hugging Face2025-06-02 更新2025-10-18 收录
下载链接：
https://hf-mirror.com/datasets/illuin-conteb/squad-conteb-eval
下载链接
链接失效反馈官方服务：
资源简介：
ConTEB - SQuAD是一个用于评估上下文嵌入模型能力的基准数据集，它源自广泛使用的SQuAD数据集。该数据集包含了2067个原始文档，从中提取了17562个片段，并对应有2067个查询。每个片段的平均标记数为19.1。数据集结构包括文档信息和查询信息，其中文档信息包含片段ID和片段文本，查询信息包含查询文本、相关答案和对应的片段ID。
This dataset, ConTEB - SQuAD, is designed to evaluate the capabilities of contextual embedding models. It is derived from the widely used SQuAD dataset and includes 2067 original documents, 17562 extracted chunks, and 2067 queries, with an average of 19.1 tokens per chunk. The dataset structure consists of document information containing chunk IDs and chunk text, and query information containing the query text, relevant answers, and the corresponding chunk ID.
提供机构：
illuin-conteb
