←back to thread

Something weird is happening with LLMs and chess

(dynomight.substack.com)
688 points crescit_eundo | 7 comments | 14 Nov 24 17:05 UTC | HN request time: 0.652s | source | bottom
Show context
azeirah ◴[14 Nov 24 22:43 UTC] No.42141993[source]▶
Maybe I'm really stupid... but perhaps if we want really intelligent models we need to stop tokenizing at all? We're literally limiting what a model can see and how it percieves the world by limiting the structure of the information streams that come into the model from the very beginning.

I know working with raw bits or bytes is slower, but it should be relatively cheap and easy to at least falsify this hypothesis that many huge issues might be due to tokenization problems but... yeah.

Surprised I don't see more research into radicaly different tokenization.

replies(14): >>42142033 #>>42142384 #>>42143197 #>>42143338 #>>42143381 #>>42144059 #>>42144207 #>>42144582 #>>42144600 #>>42145725 #>>42146419 #>>42146444 #>>42149355 #>>42151016 #
aithrowawaycomm ◴[14 Nov 24 23:25 UTC] No.42142384[source]▶
FWIW I think most of the "tokenization problems" are in fact reasoning problems being falsely blamed on a minor technical thing when the issue is much more profound.

E.g. I still see people claiming that LLMs are bad at basic counting because of tokenization, but the same LLM counts perfectly well if you use chain-of-thought prompting. So it can't be explained by tokenization! The problem is reasoning: the LLM needs a human to tell it that a counting problem can be accurately solved if they go step-by-step. Without this assistance the LLM is likely to simply guess.

replies(6): >>42142733 #>>42142807 #>>42143239 #>>42143800 #>>42144596 #>>42146428 #
1. Der_Einzige ◴[15 Nov 24 00:29 UTC] No.42142807[source]▶
I’m the one who will fight you including with peer reviewed papers indicating that it is in fact due to tokenization. I’m too tired but will edit this for later, so take this as my bookmark to remind me to respond.
replies(4): >>42142884 #>>42144506 #>>42145678 #>>42147347 #
2. aithrowawaycomm ◴[15 Nov 24 00:40 UTC] No.42142884[source]▶
I am aware of errors in computations that can be fixed by better tokenization (e.g. long addition works better tokenizing right-left rather than L-R). But I am talking about counting, and talking about counting words, not characters. I don’t think tokenization explains why LLMs tend to fail at this without CoT prompting. I really think the answer is computational complexity: counting is simply too hard for transformers unless you use CoT. https://arxiv.org/abs/2310.07923
replies(1): >>42143144 #
3. cma ◴[15 Nov 24 01:34 UTC] No.42143144[source]▶
Words vs characters is a similar problem, since tokens can be less one word, multiple words, or multiple words and a partial word, or words with non-word punctuation like a sentence ending period.
4. Jensson ◴[15 Nov 24 06:55 UTC] No.42144506[source]▶
We know there are narrow solutions to these problems, that was never the argument that the specific narrow task is impossible to solve.

The discussion is about general intelligence, the model isn't able to do a task that it can do simply because it chooses the wrong strategy, that is a problem of lack of generalization and not a problem of tokenization. Being able to choose the right strategy is core to general intelligence, altering input data to make it easier for the model to find the right solution to specific questions does not help it become more general, you just shift what narrow problems it is good at.

5. azeirah ◴[15 Nov 24 10:46 UTC] No.42145678[source]▶
I strongly believe that the problem isn't that tokenization isn't the underlying problem, it's that, let's say bit-by-bit tokenization is too expensive to run at the scales things are currently being ran at (openai, claude etc)
replies(1): >>42150150 #
6. pmarreck ◴[15 Nov 24 14:43 UTC] No.42147347[source]▶
My intuition says that tokenization is a factor especially if it splits up individual move descriptions differently from other LLM's

If you think about how our brains handle this data input, it absolutely does not split them up between the letter and the number, although the presence of both the letter and number together would trigger the same 2 tokens I would think

7. int_19h ◴[15 Nov 24 19:39 UTC] No.42150150[source]▶
It's not just a current thing, either. Tokenization basically lets you have a model with a larger input context than you'd otherwise have for the given resource constraints. So any gains from feeding the characters in directly have to be greater than this advantage. And for CoT especially - which we know produces significant improvements in most tasks - you want large context.