zlyngkhoi/bbq_unlearning_v3_train_test_split

该数据集包含了问题、答案以及相关上下文信息。每个样本都包括了问题的索引、极性、条件、类别等属性，并且有三个可能的答案。此外，还包含了额外的元数据，如数据来源、刻板印象群体、子类别和版本。数据集分为训练集和测试集两部分。

The dataset includes questions, answers, and related context information. Each sample contains attributes such as question index, polarity, condition, category, and three possible answers. In addition, it includes additional metadata such as source, stereotyped groups, subcategory, and version. The dataset is divided into training and test sets.