UnQover

该数据集包含了一系列非负面问题，用于分析语言模型输出中的偏见。此外，值得注意的是，UnQover数据集并未提供“无法确定”的选项，但模型有时会返回“未知”的答案，这类答案被归类在“未知”类别中。这项任务旨在对跨语言模型进行偏见评估。

This dataset comprises a series of non-negative questions, designed for analyzing biases in the outputs of language models. It is worth noting that the UnQover dataset does not offer an "unable to determine" option, yet models occasionally return "unknown" answers, which are classified under the "unknown" category. This task is intended for bias evaluation of cross-lingual language models.