Hacker News new | ask | show | jobs
by visarga 872 days ago
everyone is using LLM as a judge to fix unbound output eval
1 comments
phillipcarter 871 days ago
No, not everyone is. It's one of the ways you can do this, though.
link
visarga 871 days ago
Maybe I exaggerated a bit, but there are many papers today going this route.
link