jigsaw_unintended_bias
收藏OpenCSG2024-07-19 更新2026-01-19 收录
下载链接：
https://opencsg.com/datasets/google/jigsaw_unintended_bias?tab=summary
下载链接
链接失效反馈官方服务：
资源简介：
Jigsaw Unintended Bias in Toxicity Classification 专注于识别和抑制网络上的不良言论。它包含英文评论文本及其毒性评分，以及多种毒性子类型和身份属性。数据规模在100万到1000万条样本之间。此数据集支持毒性预测等多属性预测任务，并采用CC0 1.0授权许可。数据来源于Kaggle竞赛，由众包方式进行标注。
Jigsaw Unintended Bias in Toxicity Classification focuses on identifying and mitigating harmful online speech. It contains English comment texts, their corresponding toxicity scores, multiple toxicity subtypes and identity attributes. The dataset contains between 1 million and 10 million samples. It supports multi-attribute prediction tasks such as toxicity prediction, and is licensed under CC0 1.0. The dataset is sourced from a Kaggle competition and annotated via crowdsourcing.
提供机构：
google
创建时间：
2024-07-19
