News for hackers. Stuff that matters.

A comment on RLHF Book. Parent comment

Re:RLHF Book by bryan0 on Saturday February 01, 2025 @10:37PM (#42905552)
Yes but these were steps were not used in R1-zero where its reasoning capabilities were trained.
[ Parent | on HN ]
Re:RLHF Book by littlestymaar on Sunday February 02, 2025 @05:28AM (#42907748)
And as a result R1-zero is way too crude to be used directly, which is a good indication that it remains relevant.
[ Parent | on HN ]