AlignmentResearch/Llama3Jailbreaks
收藏Hugging Face2025-02-12 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/AlignmentResearch/Llama3Jailbreaks
下载链接
链接失效反馈官方服务：
资源简介：
这是一个文本分类数据集，包含三个配置：默认配置、负样本配置和正样本配置。每个配置都有训练集和验证集。数据集的特征包括完成情况、指导、回答提示、内容、分类标签、代理分类标签、生成目标和代理生成目标。分类标签和代理分类标签用于标识文本内容是否为良性或有害。
This is a text classification dataset with three configurations: default, negative samples, and positive samples. Each configuration includes a training set and a validation set. The dataset features include completion, instructions, answer prompt, content, classification label, proxy classification label, generation target, and proxy generation target. The classification label and proxy classification label are used to identify whether the text content is benign or harmful.
提供机构：
AlignmentResearch
