marcov/adversarial_qa_dbert_promptsource

该数据集包含多个特征，如id、标题、上下文、问题、答案等，其中答案是一个序列，包含文本和答案起始位置两个子特征。此外，数据集还包含元数据、模板名称、模板、渲染输入和渲染输出等特征。数据集分为训练集、验证集和测试集三个部分，分别包含50000、5000和1000个样本。数据集的下载大小为24367343字节，总大小为115463872.0字节。

The dataset contains multiple features such as id, title, context, question, answers, etc., where answers is a sequence containing two sub-features: text and answer_start. Additionally, the dataset includes metadata, template_name, template, rendered_input, and rendered_output. The dataset is divided into train, validation, and test sets, containing 50,000, 5,000, and 1,000 samples respectively. The download size of the dataset is 24,367,343 bytes, and the total size is 115,463,872.0 bytes.