LLMs can't self-correct in reasoning tasks, DeepMind study finds

Re:LLMs can't self-correct in reasoning tasks, DeepMind stud... by YeGoblynQueenne on Monday October 09, 2023 @04:44PM (#37825265)

Hang on. I've watched these trainwrecks for a while. The people who usually don't let details get in their way are the ones who claim that LLMs think even though these same people can't explain what thinking means.

The onus [1] should go the other way.

____________

[1] Today's conversation-derailing trivium: "onus" means ass in Greek. As in donkey.

https://en.wikipedia.org/wiki/Onager

[ Parent | on HN ]

Re:LLMs can't self-correct in reasoning tasks, DeepMind stud... by smokel on Monday October 09, 2023 @04:55PM (#37825388)

I'm happy to be on the agnostic front.

A colleague at my first job was working on some software that wrapped around other software, and he dubbed it "Burrito". When his project turned out not to be as useful as expected, he was pleasantly surprised that he could also use the original meaning of the word burrito, being "little donkey" in Spanish [1].

[1] https://en.wikipedia.org/wiki/Burrito

[ Parent | on HN ]

Re:LLMs can't self-correct in reasoning tasks, DeepMind stud... by YeGoblynQueenne on Monday October 09, 2023 @05:09PM (#37825514)

[ Parent | on HN ]

Re:LLMs can't self-correct in reasoning tasks, DeepMind stud... by ryanklee on Monday October 09, 2023 @04:53PM (#37825358)

It happens on both sides in equal measure.

[ Parent | on HN ]