Possible Bug Or Glitch

API
1

I keep getting an alert that I’m typing toxic content or whatever the term is, but I’m really not. I didn’t take a picture of the other time I experienced this, but for this one, I asked a seemingly basic question and it threw up this error (my question was about different AI/ML models):

image
image1125×2436 78.3 KB

The other scenario, I was experimenting with the tweet feature where it can detect the emotion of the tweet. I typed in something like “I despise the newest Marvel movie!” and it threw up that same toxic error that’s in the screenshot. It’s possible there is something wrong on my end as the Marvel movie one was my first time experimenting and the model question was my second time. If this is not on my end, then it might be a bug. If this is my fault, please explain what I’m doing wrong.

Thank you so much for your time! Have a wonderful day!

PS I don’t actually dislike Marvel movies (don’t worry lol)

3 Likes
2

You can click “report issue” to report when you think the content filter made a mistake.

I agree that the screenshotted example shouldn’t trigger the warning, but it’s subjective whether or not “I despise _” should trigger it or not.

In any case, the content warning doesn’t actually impact your completions or use of the playground in any way. Again, if you feel the warning was triggered incorrectly, you can use “report issue.”

1 Like
3

Also the response it generated could actually have generated toxic content. We wont know since he never showed us the response.