OpenAI: support for Reinforcement Fine-tuning available to verified orgs (opens in new tab)

(twitter.com)

1 pointsjustanotheratom1y ago1 comments

1 comments

1 comments · 1 top-level

my question for anyone who knows:

Between SFT, DPO, and RFT, - when to use which? - can we mix and match? e.g, first SFT, then DPO.

j / k navigate · click thread line to collapse