(arxiv.org)

47 points Jimmc414 | 1 comments | 30 Apr 24 19:20 UTC | HN request time: 0.208s | source

1. xianshou ◴[30 Apr 24 21:09 UTC] No.40216381[source]▶

I was going to knock this for incrementally improving performance while massively increasing costs, but it's actually 7x less expensive. Not bad.

Replacing Judges with Juries: Evaluating LLM Generations with a Panel of Models