five

Dataset for Interactive Profiling Narrative (IPN) with Toxicity Tolerance Score of each individual user to different categories of toxicity

收藏
NIAID Data Ecosystem2026-05-02 收录
下载链接:
https://zenodo.org/record/14196542
下载链接
链接失效反馈
官方服务:
资源简介:
he dataset is the collection of the results of the Interactive Profiling Narrative that was developed as a part of the project Listener Aware Content Detoxification.It contains the toxicity tolerance scores of each individual user to different categories.High scores indicate that the user is extremely sensitive to the particular category where as low scores indicate that the user isn't that triggered by the category.Medium scores show moderate tolerance. Columns:1.UserId: The unique id given to each user (helps preserve anonymity).2.Race: The user's toxicity tolerance to the category race. High score means racist comment trigger them. 3.Sex: The user's toxicity tolerance to the category sexuality. High scores indicate sensitivity to comments on sexual identity.4.Body_image: The user's sensitivity to comments regarding body image. High scores indicate that remarks on body appearance strongly affect the user.5.Disability:The user's sensitivity to comments about disabilities. A high score means the user is highly sensitive to potentially ableist remarks.6.Religion_culture: The user's sensitivity to content involving religion or culture. High scores show the user is easily triggered by insensitive comments about religious or cultural aspects.7.Physical_abuse: User's sensitivity to comments regarding physical abuse. High scores indicate the user is strongly affected by remarks on physical abuse.8.Mental_health: This measures the user's sensitivity to comments on mental health. High scores suggest that the user is particularly affected by comments stigmatizing mental health issues.9.Politics:The user's sensitivity to political content. High scores indicate that political discussions or comments are likely to evoke a strong reaction in the user
创建时间:
2024-11-21
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作