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
Actually this part does seem in recent research to be encoded in LLMs at an abstract level in a linear representation...
https://arxiv.org/abs/2310.06824
There are no comments on this comment yet.
Actually this part does seem in recent research to be encoded in LLMs at an abstract level in a linear representation...
https://arxiv.org/abs/2310.06824