>>7891288 lb
>>7894304 lb
>>7894670 lb
Some exploration of the Google Toxicity score tool from this website:
https://www.perspectiveapi.com/#/home
Toxicity and sub-attribute annotation guidelines:
https://github.com/conversationai/conversationai.github.io/blob/master/crowdsourcing_annotation_schemes/toxicity_with_subattributes.md
Wikipedia Talk Labels: Toxicity
This shows three datasets but I'm not computer savvy enough to understand what's going on.
https://figshare.com/articles/Wikipedia_Talk_Labels_Toxicity/4563973
FAQs on Perspective:
https://support.perspectiveapi.com/s/article/perspective-faqs
I think this is part of the censorship algorithm used in social media and MSM, but I can't find a direct connection other than this screencap that shows users of the Perspective API.
https://www.perspectiveapi.com/#/home