marcov/trivia_qa_unfiltered_promptsource

该数据集包含多个特征字段，如问题、问题ID、问题来源、实体页面、搜索结果和答案等。实体页面和搜索结果字段包含多个子字段，如文档来源、文件名、标题、维基上下文、描述、排名、URL和搜索上下文等。答案字段包含别名、归一化别名、匹配的维基实体名称、归一化匹配的维基实体名称、归一化值、类型和值等子字段。数据集分为训练集、验证集和测试集，分别包含438110、56565和43328个示例。数据集的总下载大小为83441044418字节，总大小为143441617765字节。

The dataset includes multiple feature fields such as question, question ID, question source, entity pages, search results, and answer. The entity pages and search results fields contain subfields like document source, filename, title, wiki context, description, rank, URL, and search context. The answer field includes subfields like aliases, normalized aliases, matched wiki entity name, normalized matched wiki entity name, normalized value, type, and value. The dataset is divided into training, validation, and test sets, containing 438110, 56565, and 43328 examples respectively. The total download size of the dataset is 83441044418 bytes, and the total size is 143441617765 bytes.