marcov/narrativeqa_promptsource

该数据集包含多个特征，如document、question、answers等，每个特征下又有多个子特征。例如，document特征包含id、kind、url、file_size、word_count、start、end、summary和text等子特征。question特征包含text和tokens子特征。answers特征是一个列表，包含text和tokens子特征。此外，数据集还包含template_name、template、rendered_input和rendered_output等特征。数据集分为train、test和validation三个部分，每个部分都有相应的字节数和示例数。

The dataset contains multiple features such as document, question, answers, etc., each with multiple sub-features. For example, the document feature includes sub-features like id, kind, url, file_size, word_count, start, end, summary, and text. The question feature includes text and tokens sub-features. The answers feature is a list containing text and tokens sub-features. Additionally, the dataset includes features like template_name, template, rendered_input, and rendered_output. The dataset is divided into train, test, and validation parts, each with corresponding byte sizes and example counts.