mteb/llm-eval-humaneval
收藏资源简介:
该数据集是一个用于信息检索或文本匹配任务的数据集,包含三个主要配置:1) corpus配置:包含158个文档,每个文档有标题和文本字段,代表语料库内容;2) queries配置:包含100个查询,每个查询有文本字段,代表用户查询;3) default配置:包含100个测试样本,每个样本关联查询ID、语料库ID和分数(整数类型),用于评估查询与文档的相关性。数据集总大小约为65KB,适用于训练或测试检索模型。
This dataset is designed for information retrieval or text matching tasks, consisting of three main configurations: 1) corpus configuration: includes 158 documents, each with title and text fields, representing the corpus content; 2) queries configuration: includes 100 queries, each with a text field, representing user queries; 3) default configuration: includes 100 test samples, each linking a query ID, corpus ID, and a score (integer type), used to evaluate the relevance between queries and documents. The total dataset size is approximately 65KB, suitable for training or testing retrieval models.



