FRESH

Hacker News

Watching a language model think before it speaks

24 points by ninjahawk1

by mherdelight

1 subcomments

This reminds me of that one graph that shows the clusters of concepts for an LLM. Very interesting.

by j_bum

1 subcomments

I’d love to see a recording of this in action for a complicated question.
This article also reads with many LLM-isms, so I can’t tell if a human actually produced it.

0 subcomment