Every story and comment from the Hacker News API, in a layout borrowed from 2001.
A comment on Infrastructure set-up & open-source scripts to train a 70B model from bare metal. Parent comment
Cool stuff! Does this do RLHF or just pretraining? If the latter, how did you manage to beat GPT 4?
There are no comments on this comment yet.
Cool stuff! Does this do RLHF or just pretraining? If the latter, how did you manage to beat GPT 4?