Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

“Next-token predictor” is the wrong mental model for LLMs

gmcgoldr.github.io

Post reply on HN
Print view View original

341–344 of 344 posts

Re: “Next-token predictor” is the wrong mental model for LLMs

#341

Post by hippietrail »

mwkaufma wrote: post #16
valleyer wrote: post #11

Earlier quoted context omitted.

During training, certain tokens are more likely to lead to a lower loss function value, which is how you "win" the game of LLM output.

So, next-token predictors

I keep seeing this stochastically parroted here. Perhaps by humans but how knows?

Re: “Next-token predictor” is the wrong mental model for LLMs

#342

Post by hippietrail »

polotics wrote: post #4

yep "next-embedding" predictor is more correct, and not just at the end but through the layers, and folding back dimensions into that one next token is one small final step, and next-embedding could be named "next-meaning" as well, and we're getting there... this sentence above would made a longer article if I bothered to so blog as is being blogged here

Exactly. There's a widespread misconception that it works on tokens all the way through. Tokens are only at the input and output edges. All the internal transformation is in the many-dimensional tensors variously described as "magic" or "not magic" or "black box", or hand-waved away as "various mathematical operations".

Re: “Next-token predictor” is the wrong mental model for LLMs

#344

Post by danielmarkbruce »

hippietrail wrote: post #339
danielmarkbruce wrote: post #43

The biggest problem is the word "predictor". Once you get into post training with RLHF and RLVR, it simply isn't doing that. It is not predicting anything. It's producing tokens, but it isn't predicting them. The chess analogy in the post is a good one - it's closer to searching for a set of moves that give a result than predict. It's search for a set of ideas, represented as locations in very high dimensional space,…

"Guessing" is more accurate than "predicting". It takes educated guesses.

It's closer to strategizing once you've run a model through RL. It's optimized to take steps which will lead it to a good outcome down the road.
Post reply on HN