[andy12_]

The automated AI pipeline also had an automatic grading model to try to reduce false positives.

But anyway, my point was that in that case the prompt involved was indeed pretty much "hey, ChatGPT, solve an unsolved problem, thanks."

[YeGoblynQueenne]

Yes, but there's no way to tell how many times the prompt was tried before a proof was found; and that only stopped when a human said it could stop.