Skip to content

Comment on My finetuned models beat OpenAI's GPT-4parent

Comments

Also a lot of the time the dates are wrong seems to be due to only having those formats, which does make me wonder again how do fine tuned get this right unless they have been fine tuned using eval data...

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.