huyhoangdinhcong/guardrails-dataset-full

该数据集包含文本字段和多个标签字段，用于分类文本是否包含性、仇恨言论、政治、暴力或有害内容。数据集分为训练集和多个验证集，分别针对不同的标签类型。数据集总大小约为113GB，下载大小约为59GB。The dataset includes a text field and multiple label fields for classifying whether the text contains sexual content, hate speech, politics, violence, or harmfulness. The dataset is split into a training set and several validation sets, each targeting different label types. The total size of the dataset is approximately 113GB, with a download size of about 59GB.