LanguageShades/BiasShadesBaseEval_mistralai_Mistral-7B-v0.1

该数据集包含多种语言的偏见句子及其相关信息，如偏见模板、是否为表达式、评论等。每个语言的偏见句子都通过Mistral-7B模型生成了token和logprob序列。数据集包含一个测试集，大小为1135452字节，包含210个样本。

This dataset contains biased sentences in multiple languages along with related information such as bias templates, whether they are expressions, comments, etc. For each language, the biased sentences have been processed by the Mistral-7B model to generate token and logprob sequences. The dataset includes a test set with a size of 1135452 bytes, containing 210 samples.