A Holistic Approach to Undesired Content Detection in the Real World

查看更多>>作者:Markov Todor;Zhang Chong;Agarwal Sandhini;Eloundou Tyna;Lee Teddy;Adler Steven;Jiang Angela;Weng Lilian
  • 创建日期:2024-04-11
  • 发布日期:2024-04-11
  • 最新更新日期:2022-10-01
简介:
查看更多>>We present a holistic approach to building a robust and useful natural language classification system for real-world content moderation. The success of such a system relies on a chain of carefully designed and executed steps, including the design of content taxonomies and labeling instructions, data quality control, an active learning pipeline to capture rare events, and a variety of methods to make the model robust and to avoid overfitting. Our moderation system is trained to detect a broad set of categories of undesired content, including sexual content, hateful content, violence, self-harm, and harassment. This approach generalizes to a wide range of different content taxonomies and can be used to create high-quality content classifiers that outperform off-the-shelf models.
DOI10.48550/arxiv.2208.03274
语种:
引用及导出
数据仓储
  • ScienceDB

  • 浏览量0
  • 点赞量0
  • 收藏量0
  • 引用量0
版本
V2 在线更新2024-04-11 00:00:00
V1 在线发布2022-11-01 00:00:00
数据集推荐