AlignmentResearch/JailbreakInputs
收藏Hugging Face2025-03-11 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/AlignmentResearch/JailbreakInputs
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含文本内容以及与之相关的标签信息，用于分类文本内容是否为良性（Benign）或有害（Harmful）。数据集分为训练集和验证集，有三个不同的配置：default、neg和pos，每个配置都有相应的数据文件路径。
The dataset contains text content along with associated label information, used for classifying whether the text content is Benign or Harmful. The dataset is split into training and validation sets, with three different configurations: default, neg, and pos, each with corresponding data file paths.
提供机构：
AlignmentResearch
