Not sure where I implied they are "raw base models" and not sure what "various adjustments" means here or how "ton of RLHF" contradicts anything. If we look at research for open source models, "adjustments" usually come in the form of efficiency gains which directly contributes to the ability to scale or synthetic data pipelines to increase the dataset and increasing the context window.
replyrlhf = reinforcement learning from human feedback
reply(had to look it up)
I think it's more RLVR (reinforcement learning from verified rewards). The RLHF is just to align models to human preferences, meaning to behave nice.
replyWhat makes you say that
replyMore RLHF is in fact scaling.
replyYes, but not in the “dump another chunk of all written language in the bucket and stir”-sense which is what bitter lesson became synonymous with.
replyThat may not be the intent of the original article, but over the past few years that’s what the phrase turned into.
GPT-6 is supposed to be using a much larger base model that just finished pretraining so the "dump another chunk of all written language" approach is still going strong.
replyWhere did they get a second copy of all books, written/audio/visual media, and the web?
replyCouch cushions?
Modern pretraining also consists of expensive human-led specialized task creation and grading loops. Synthetic generation and distillation from previous models is another input for training. I wonder how much new text contributes beyond keeping knowledge up-to-date.
replyThe measure is to see if the results scale, not just the rumored attempts at building such a model, as o3 taught us.
replyWhere are they getting new data from at this point? Didn't they already read the entire internet?
replyCustom sets and mining their own users, as every lab does
replyThe bitter lesson just means “compute scaling beats hand-tuned architectures in the long run”.
replyAs GP said. More RLHF is in fact the bitter lesson.
deleted
replyNot just RLHF but also RLVR, and isn't that the litter lesson though?
replyMy sense of the Sutton Dwarkesh interview was that he was calling out that he didn't mean just longer datasets, but rather learning through exploration and that's exactly RL.
They just need more contact with reality. That's what RL is right? Contact with narrow subsets of reality.
replyRLHF is an increasingly small part of training though? From what I understand most of the capability gain is in RLVR
replyNah, the last few generations have more RLVR in the data mix. Which is more CPU intensive and very much amenable to the bitter lesson as you can reduce the loss by doing more rollouts in your tool environment.
replyThe scaling with reasoning models is more and more with things like verifiable rewards (coding and math), in line with bitter lesson and also Sutton invented lots of modern RL.
reply