Re:Are popular toxicity models simply profanity detectors?
by selestify on Tuesday January 25, 2022 @12:13AM
(#30068358)
I searched for “tourism” and “Kristen Gray” hoping to get a tweet like [1], but alas the results were actually reasonable :)
[1] https://twitter.com/celesteperez___/status/13508599618452070...
Re:Are popular toxicity models simply profanity detectors?
by npilk on Tuesday January 25, 2022 @12:38AM
(#30068523)
Ah - the oracle only looks at recent tweets, so it may not be that helpful for past controversies...
In general I've found the results to be much better for subjects that a lot of people are currently tweeting about.
Re:Are popular toxicity models simply profanity detectors?
by sodality2 on Tuesday January 25, 2022 @12:39AM
(#30068526)
I tried "hacker news" and got "Hacker News is an internet oasis". :D cool!!
Re:Are popular toxicity models simply profanity detectors?
by newbytuby on Monday January 24, 2022 @11:34PM
(#30068052)
pretty cool idea! would be interesting to see how the model's selection changes with more specific data
You can see that my model is basically just filtering for profanity as an indicator of "strong emotion", which makes sense. But it's interesting that postive profanity seems to be such a thorny problem, at least for Perspective.