Re:LLMs can't self-correct in reasoning tasks, DeepMind stud...
by soco on Tuesday October 10, 2023 @03:51AM
(#37829534)
And I assume there's no understanding of what "real understanding" means. In this so limited context "real understanding" looks like "did what I wanted without me really explaining what I wanted". Although in real life I seldom know from the start what I really want either and takes me a few iterations to clarify it even to myself - so "real understanding" seems to be something even I am missing (and I shouldn't blame poor AI for lacking it). Or maybe I'm an AI myself just unaware of it?
[ Parent
| on HN ]
There are no comments on this comment yet.