Hacker News new | ask | show | jobs
by scosman 714 days ago
Re: 7b models vs gpt-3.5, I’m guessing different fine tuning parameters can account for the difference. The OpenAI fine tuning is a black box.