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

How LLMs work

0xkato.xyz

Post reply on HN
Print view View original

201–210 of 293 posts

Re: How LLMs work

#201

Post by kzrdude »

chris_money202 wrote: post #190
miki123211 wrote: post #166

There's one thing I wish people understood about LLMs, and it doesn't really have anything to do with what's inside the neural network part. It's the fact that LLMs can only write in one direction — forward. When you are writing an essay and realize midway through a sentence that what you've written doesn't make sense, you go back and edit. An LLM can't do that, the only thing it can do is keep on generating. Because…

In terms of our brains though we can only think forward as well (if forward is time). Our brain in the future says something we did in the past was wrong (part of the sentence we wrote) and that informs our body (the agent) to go back and fix it

I think that's why pen and paper is such a good tool for thinking. :)

Re: How LLMs work

#202

Post by wizzwizz4 »

zemo wrote: post #133
mattnewton wrote: post #28

Earlier quoted context omitted.

I don’t disagree with your conclusion that this is likely ai rewritten, but I do find it strange that you say “normal people don’t write like this” when it is mimicking how people write, and using patterns I have seen people write. I think models are at the point where style is not really reliable as an indicator anymore.

It’s not a model of an author, it’s a model of documents. That’s not the same thing.

No, but sufficiently-advanced overfitting would lead to to the model keeping track of an author stylistic profile, in the same way it keeps track of the plot of a story it's writing (i.e., badly, but well enough that you have to pay attention to notice that something is wrong).

Re: How LLMs work

#203

Post by layla5alive »

Borealid wrote: post #148
layla5alive wrote: post #113

Earlier quoted context omitted.

Lol, the bird did not 'fly' - it just flapped its wings and generated lift!

No. The how is relevant here because it leads to understanding of the resulting behavior. If you train the LLM on a corpus that shows people saying the sky is red, you get an LLM that is predisposed to say the sky is red. This is true even if it's also trained on all of the science that explains how and why the sky is blue. If it were to "figure out" or "reason", it would not have such a predisposition to emit "red"…

You can also teach humans science and math and then they can be trained by a cult to not use any of that reasoning when emitting canned responses that they were rewarded by the cult for internalizing during their training. "Fake News!"

You're caught up on the mechanics of token processing (floating point matrix ALU math) and ignoring the context that p(next token) as a function being "computed" is doing so over a trillion parameters. You can poorly train a model, sure, but assuming you don't indoctrinate it too much, properties like cognition emerge - it learns to reason; why? Reasoning is more efficient and compact than memorizing answers.

Re: How LLMs work

#204

Post by swyx »

xnx wrote: post #189
ekunazanu wrote: post #76

Earlier quoted context omitted.

> This was openai’s entire breakthrough. Making this particular model architecture larger leads to emergent capabilities Basically, the bitter lesson: https://www.cs.utexas.edu/~eunsol/courses/data/bitter_lesson...

Isn't the bitter lesson basically the same as "The Unreasonable Effectiveness of Data" from 2009?

not exactly, bitter lesson is one meta-level up from "scale eats everything". this is a common misunderstanding of bitter lesson that rich sutton has been fighting ever since the thing was written. in rich's own words[1], the modern summary is

> Don’t be distracted by human knowledge, as AI has been historically.

> Instead focus on methods for creating knowledge that scale with computation, like search and learning.

so the lesson is choose methods that scale with computation, not just that blindly scaling up anything (data, params, people, whatever) works, it is choosing the right x axis and the right scaling laws consistently wins out in the long run despite short term wins from other methods.

1: https://x.com/RichardSSutton/status/2056419165502935198

Re: How LLMs work

#205

Post by swyx »

gobdovan wrote: post #54
achrono wrote: post #31

Earlier quoted context omitted.

How do we know that today's frontier models are merely scaled up versions of that? Genuine question, since the labs have narrowed what they share over the years to now almost nothing, in terms of how the model was trained and how it works under the hood.

DeepSeek research: - V3 https://arxiv.org/abs/2412.19437 - V2 https://arxiv.org/abs/2405.04434 - R1 https://arxiv.org/abs/2501.12948 (RL applied to ML models was well-known beforehand, but they show it in the open, at scale, on big models) Then, there's the incentive analysis. If you can see that these models empirically get better with scale, why would you swap the main architecture? Those events will be pretty rare…

> If you can see that these models empirically get better with scale, why would you swap the main architecture? Those events will be pretty rare

c.f. hardware lotter https://arxiv.org/abs/2009.06489

Re: How LLMs work

#206

Post by redox99 »

sirsinsalot wrote: post #184
redox99 wrote: post #134

Earlier quoted context omitted.

Almost everything in ML is like that. It seems so obvious in hindsight. It's maybe what I love most. Residual connections are so simple, so obvious and so vital. Yet nobody came up with them until 2015?

I suspect it was considered many times, but the sheer computation scale would make it feel like obscene brute force. It feels like the right shape but too wild to think about implementing. I think as time went on, and hardware got better, it seemed more reasonable to actually think about a viable implementation of what I think was a widespread intuition anyone in ML had that everything's context is everything. It jus…

People definitely wanted to train deep networks before, but didn't know how. They evdn tried things like training layers independently.

I don't think it was intuitive to anyone back then, the vanishing gradient problem was a big deal since the dawn of NNs. I'm not sure what you mean by sheer computation, residuals allow you to have deep networks instead of shallow and wide ones. You can have equivalent parameter count.

Re: How LLMs work

#207

Post by spaceisballer »

Solid read especially for someone not in this field. While everything I’ve learned about LLMs has been pretty interesting, all I can say for sure is that I’m more and more skeptical about wide scale adoption. Consumers are being pushed almost to the level of coercion to utilize LLMs. Especially in the case of the government, who in most cases will get a free pass for a year to help build that addiction before the real bill comes due. I’m sure people will object to LLMs being considered statistical bullshit generators, but I cant stop seeing that they just generate bullshit. Bullshit can be believable, sometimes bullshit is truthful. But the public is generally accepting that LLMs churn out truth, and they trust them blindly. I truly don’t see the upside, and the argument being pushed by the dealers is that we need to keep using. The real breakthrough is around the corner. It’s my believe that the tech bros are the new robber barons, they’ve moved on from crypto and NFTs and found something that is pretty impressive but a far cry from the GenAI that wish for. I just feel that technology was supposed to free us and also connect us, but instead it’s all things addictive and consuming.

Re: How LLMs work

#208

Post by swyx »

miki123211 wrote: post #166

There's one thing I wish people understood about LLMs, and it doesn't really have anything to do with what's inside the neural network part. It's the fact that LLMs can only write in one direction — forward. When you are writing an essay and realize midway through a sentence that what you've written doesn't make sense, you go back and edit. An LLM can't do that, the only thing it can do is keep on generating. Because…

agree reasoning fixed a lot of it. but the inherent path dependence of autoregression is one reason i was excited for text diffusion models (https://www.youtube.com/watch?v=r305-aQTaU0)

instead of going left to right, even with a scratchpad, maybe you start with a rough shape of the big picture all at once, and then you iteratively resolve and things come into focus.

mercury (https://www.youtube.com/watch?v=2fDBeMu6xjk) seems to have made the most progress here, which is not saying a ton but is not nothing. i do think it is telling that of the big labs, only GDM has made any meaningful bet on text diffusion. you can bet your ass all of them have evaluated it for a source of alpha.

Re: How LLMs work

#209

Post by 0xbadcafebee »

rfv6723 wrote: post #62
ctolsen wrote: post #38

Earlier quoted context omitted.

No, it’s definitely not what a human brain is. That makes very little sense. The ways we interact with language (and thus conceptual memory) is completely and fundamentally different.

Is it different though? If we look beyond written languages which are late inventions of human civilization, oral languages are continuous and build with blocks not words. Chomskyan school misled the entire field of linguistics for decades by ignoring spoken languages.

Chomsky did the opposite of what you're saying. He didn't ignore spoken language. He said that human vocalization is independent of language, and that the way our brains can manipulate and use sound (a cognitive capability, not specifically an aural one) is the fundamental differentiator that allows us to make compound ideas, and our specific use of language is a byproduct.

Example: a programming language's capability to produce complex software does not come from some inherent quality of language. It comes from binary. 0's and 1's, representing basic logic, and that being built on top of with an abstract "tool" called a language. If the binary logic didn't work, the language wouldn't do anything.

A dolphin can make sounds, and technically has a language, but they can't manipulate or recursively compound concepts (as far as we can tell) in order to create modified ideas. If they could, they probably would have come up with vastly more advanced fishing methods than the (admittedly novel) ones they have now.

Re: How LLMs work

#210

Post by Izkata »

baq wrote: post #41
ai_slop_hater wrote: post #36

Earlier quoted context omitted.

Human brain capabilities are truly amazing, imagine if people didn’t treat their children as if they are stupid and didn’t constantly lie to them, because kids are stupid right, they wouldn’t understand. What heights could be reached.

We don’t treat children like they’re stupid, we treat children like they’re children. A stupid adult is treated very differently than any child. Adults are expected to have their world models approximately correct in terms of physical environment so they won’t accidentally kill themselves by falling off a cliff; then there are the social norms which adults are expected to conform to so everyone is kinda predictable t…

Another example, my parents taught me to read at about 4 years old. When I started kindergarten (the year before 1st grade in the US), the teachers and principal didn't believe I could read and I had to prove it by reading a book to them I'd never seen before.

I think they're right that kids (at least in the US) are generally treated as less capable than they are, and it ends up slightly delaying their development.

Post reply on HN