This is why using other LLMs as scorers for benchmarks and evaluations is such a... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

embedding-shape 18 days ago | parent | context | favorite | on: Three sites made 215,128 “best software” pages for...

This is why using other LLMs as scorers for benchmarks and evaluations is such a bad idea, they'll have preferences you can't anticipate and won't understand immediately.

Terr_ 18 days ago [–]

The idea that that LLM reliability or bias can be solved with more LLM is... infuriatingly persistent.

nemonemo 18 days ago | [–]

Isn't this basically the mythical man-months LLM edition?

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact