ToxicityPrompts/RTP-LX

该数据集包含多个特征，包括提示、完成、索引、地区、方言、提示注释、提示原始注释、完成注释、完成原始注释和语言。注释部分包括偏见、身份攻击、侮辱、微侵犯、自残、性内容、毒性和暴力等子特征。数据集分为测试集，包含30253个样本，总大小为65390470字节。The dataset includes multiple features such as Prompt, Completion, Index, Locale, Dialect, PromptAnnotations, PromptAnnotationsRaw, CompletionAnnotations, CompletionAnnotationsRaw, and language. The annotations include sub-features like Bias, IdentityAttack, Insult, Microaggression, SelfHarm, SexualContent, Toxicity, and Violence. The dataset is split into a test set with 30253 samples and a total size of 65390470 bytes.