marcov/quartz_promptsource

该数据集包含多个字段，用于存储问题和答案的相关信息。主要字段包括问题的唯一标识符（id）、问题文本（question）、选项（choices，包含选项文本和标签）、正确答案（answerKey）、相关段落（para）、段落标识符（para_id）、段落注释（para_anno，包含效果属性、原因方向字符串、效果方向字符串等）、问题注释（question_anno，包含更多效果方向、较少效果方向等）、模板名称（template_name）、模板（template）、渲染输入（rendered_input）和渲染输出（rendered_output）。数据集分为训练集、测试集和验证集，分别包含21568、6272和3072个示例。数据集的总下载大小为7886879字节，总数据集大小为35620232字节。

This dataset contains multiple fields for storing information related to questions and answers. The main fields include a unique identifier for the question (id), the question text (question), choices (containing choice text and labels), the correct answer (answerKey), related paragraphs (para), paragraph identifiers (para_id), paragraph annotations (para_anno, containing effect properties, cause direction strings, effect direction strings, etc.), question annotations (question_anno, containing more effect direction, less effect direction, etc.), template names (template_name), templates (template), rendered inputs (rendered_input), and rendered outputs (rendered_output). The dataset is divided into training, test, and validation sets, containing 21,568, 6,272, and 3,072 examples respectively. The total download size of the dataset is 7,886,879 bytes, and the total dataset size is 35,620,232 bytes.