aieng-lab/gradiend_religion_data

这是一个包含敏感宗教词汇的模板句子的数据集，例如“犹太人”。该数据集是Wikipedia-10的过滤版本，仅包含包含源宗教的偏见敏感词汇的句子。基于掩码术语（源），从相应的偏见属性对中派生出一个关联的目标，匹配源的大小写（例如，“JEWISH”对应“CHRISTIAN”而不是“christian”）。This dataset consists of templated sentences with the masked word being sensitive to religion, e.g., *Jewish*. It is a filtered version of Wikipedia-10 containing only sentences that include a religion bias sensitive word of the `source_id` religion. Based on the masked term (`source`), an associated `target` is derived from a corresponding bias attribute pair, matching the casing of `source` (e.g., `JEWISH` gets to `CHRISTIAN` and not `christian`).