CORGI-PM
收藏资源简介:
CORGI-PM是由北京人工智能研究院等机构联合创建的中文数据集,旨在探测和缓解性别偏见。该数据集包含32,900条高质量标注的句子,涵盖性别偏见与非偏见样本。数据集通过特定的标注方案和专业背景的标注者进行创建,主要用于性别偏见的自动检测、分类和缓解。CORGI-PM的应用领域主要集中在自然语言处理中性别偏见的探测与缓解,为解决文本中的性别偏见问题提供了重要资源。
CORGI-PM is a Chinese-language dataset jointly created by institutions including the Beijing Academy of Artificial Intelligence, which aims to detect and mitigate gender bias. This dataset contains 32,900 high-quality annotated sentences covering both gender-biased and non-biased samples. It was constructed via a dedicated annotation scheme and professional annotators, and is primarily intended for the automatic detection, classification and mitigation of gender bias. The application scenarios of CORGI-PM mainly focus on the detection and mitigation of gender bias in natural language processing, providing a critical resource for addressing gender bias issues in text.




