Every story and comment from the Hacker News API, in a layout borrowed from 2001.
A comment on LLMs cannot find reasoning errors, but can correct them. Parent comment
anyway, LLMs aren't thinking. they're pattern matching and it's not doing recursion it seems.
I'd say the only way you're getting error correction is taking multiple LLMS And running them through chains and parallel construction.
There are no comments on this comment yet.
anyway, LLMs aren't thinking. they're pattern matching and it's not doing recursion it seems.
I'd say the only way you're getting error correction is taking multiple LLMS And running them through chains and parallel construction.