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

LLMs use a surprisingly simple mechanism to retrieve some stored knowledge

news.mit.edu

Post reply on HN
Print view View original

111–120 of 156 posts

Re: LLMs use a surprisingly simple mechanism to retrieve some stored knowledge

#111

Post by szundi »

visarga wrote: post #51
retrofrost wrote: post #48

This is amazing work, but to me it highlights some of the biggest problems in the current AI zeitgeist, we are not really trying to work on any neuron or ruleset that isnt much different from the perceptron thats just a sumnation function. Is it really that suprising that we just see this same structure repeated in the models. Just because feedforward topologies with single neuron steps are the easiest to train and r…

> Just because feedforward topologies with single neuron steps are the easiest to train and run on graphics cards does that really make them the actual best at accomplishing tasks? You are ignoring a mountain of papers trying all conceivable approaches to create models. It is evolution by selection, in the end transformers won.

“end”

Re: LLMs use a surprisingly simple mechanism to retrieve some stored knowledge

#112

Post by wizzwizz4 »

refulgentis wrote: post #103
foobiekr wrote: post #78

Earlier quoted context omitted.

"won" They barely work for a lot of cases (i.e., anything where accuracy matters, despite the bubble's wishful thinking). It's likely that something will sunset them in the next few years.

It seems cloyingly performative grumpy old man once you're at "it barely works and it's a bubble and blah blah" in response to a discussion about their comparative advantage (yeah, they won, and absolutely convincingly so)

That's like saying Bitcoin won cryptography.

Re: LLMs use a surprisingly simple mechanism to retrieve some stored knowledge

#113

Post by tel »

mjburgess wrote: post #46
numeri wrote: post #27

Earlier quoted context omitted.

I didn't imply that they know anything about where atoms are, I was just pointing out the sheer absurdity of that volume of data. I should make it clear that my comparison there is unfair and mostly just funny – you don't need to store every possible combination of 10 tokens, because most of them will be nonsense, so you wouldn't actually need that much storage. That being said, it's been fairly solidly proven that L…

> fairly solidly proven that LLMs aren't just lookup tables/stochastic parrots Well i'd strongly disagree. I see no evidence of this; I'm am quite well acquainted with the literature. All empirical statistical AI is just a means of approximating an empirical distribution. The problem with NLP is that there is no empirical function from text tokens to meanings; just as there is no function from sets of 2D images to a…

That suggests that no statistical method could ever recover hidden representations though. And that’s patently untrue. Taken to its greatest extreme you shouldn’t even be able to guess between two mixed distributions even when they have wildly non-overlapping ranges. Or put another way, all of statistical testing in science is flawed.

I’m not saying you believe that, but I fail to see how that situation is structurally different from what you claim. If it’s a matter of degree, how do you feel things change as the situation becomes more complex?

Re: LLMs use a surprisingly simple mechanism to retrieve some stored knowledge

#114

Post by tel »

uoaei wrote: post #86

This is the "random linear projections as memorization technique" perspective on Transformers. It's not a new idea per se, but nice to see it fleshed out. If you dig into this perspective, it does temper any claims of "cognitive behavior" quite strongly, if only because Transformers have such a large capacity for these kinds of "memories".

Do you have a reference on “random linear projections as memorization”? I know random projections quite well but haven’t seen that connection.

Re: LLMs use a surprisingly simple mechanism to retrieve some stored knowledge

#115

Post by antonvs »

visarga wrote: post #51
retrofrost wrote: post #48

This is amazing work, but to me it highlights some of the biggest problems in the current AI zeitgeist, we are not really trying to work on any neuron or ruleset that isnt much different from the perceptron thats just a sumnation function. Is it really that suprising that we just see this same structure repeated in the models. Just because feedforward topologies with single neuron steps are the easiest to train and r…

> Just because feedforward topologies with single neuron steps are the easiest to train and run on graphics cards does that really make them the actual best at accomplishing tasks? You are ignoring a mountain of papers trying all conceivable approaches to create models. It is evolution by selection, in the end transformers won.

I’d say it’s more that transformers are in the lead at the moment, for general applications. There’s no rigorous reason afaik that it should stay that way.

Re: LLMs use a surprisingly simple mechanism to retrieve some stored knowledge

#116

Post by antonvs »

erisinger wrote: post #69
Kerb_ wrote: post #62

Earlier quoted context omitted.

You can think chemotherapy is a local maxima for cancer treatment and hope medical research seeks out other options without having the resources to do it yourself. Not all of us have access to the tools and resources to start experimenting as casually as we wish we could.

Not a single one of you bigbrains used the word "maxima" correctly and it's driving me crazy.

“Maxima” sounds fancy, making it catnip for people trying to sound smart.

Re: LLMs use a surprisingly simple mechanism to retrieve some stored knowledge

#117

Post by oldpersonintx »

foobiekr wrote: post #78
visarga wrote: post #51

Earlier quoted context omitted.

> Just because feedforward topologies with single neuron steps are the easiest to train and run on graphics cards does that really make them the actual best at accomplishing tasks? You are ignoring a mountain of papers trying all conceivable approaches to create models. It is evolution by selection, in the end transformers won.

"won" They barely work for a lot of cases (i.e., anything where accuracy matters, despite the bubble's wishful thinking). It's likely that something will sunset them in the next few years.

[dead]

Re: LLMs use a surprisingly simple mechanism to retrieve some stored knowledge

#118

Post by qlk1123 »

wongarsu wrote: post #50
Animats wrote: post #30

Earlier quoted context omitted.

That's a good point. It may be possible to directly upload predicate-type info into a LLM. This could be especially useful if you need to encode tabular data. Somewhere, someone probably read this and is thinking about how to export Excel or databases to an LLM. It's encouraging to see people looking inside the black box successfully. The other big result in this area was that paper which found a representation of a…

The opposite is also exciting: build a loss function that punishes models for storing knowledge. One of the issues of current models is that they seem to favor lookup over reasoning. If we can punish models (during training) for remembering that might cause them to become better at inference and logic instead.

I believe it will add some spice to the model, but you shouldn't go too far at that direction. Any social system has a rule set, which has to be learnt and remembered, not infered.

Two exmaples. (1) grammars in natural languages. You can just see in another commenter here uses "a local maxima", and then how people react to that. I didn't even notice becuase English grammar has never been native to me. (2) Mostly, prepositions between two languages, no matter how close they are, don't have a direct mapping. The learner just has to remember it.

Re: LLMs use a surprisingly simple mechanism to retrieve some stored knowledge

#119

Post by azinman2 »

wongarsu wrote: post #50
Animats wrote: post #30

Earlier quoted context omitted.

That's a good point. It may be possible to directly upload predicate-type info into a LLM. This could be especially useful if you need to encode tabular data. Somewhere, someone probably read this and is thinking about how to export Excel or databases to an LLM. It's encouraging to see people looking inside the black box successfully. The other big result in this area was that paper which found a representation of a…

The opposite is also exciting: build a loss function that punishes models for storing knowledge. One of the issues of current models is that they seem to favor lookup over reasoning. If we can punish models (during training) for remembering that might cause them to become better at inference and logic instead.

But how to do when pre training is basically predict the next token?
Post reply on HN