Are modern LLMs closer to AGI or next word predictor? Where do they fall in this graph with 10 on x-axis being human intelligence.
Wondering if Modern LLMs like GPT4, Claude Sonnet and llama 3 are closer to human intelligence or next word predictor. Also not sure if this graph is right way to visualize it.
That's literally how llma work, they quite literally are just next word predictors. There is zero intelligence to them.
It's literally a while token is not "stop", predict next token.
It's just that they are pretty good at predicting the next token so it feels like intelligence.
So on your graph, it would be a vertical line at 0.
What is intelligence though? Maybe I'm getting through life just by being pretty good at predicting what to say or do next...
yeah yeah I've heard this argument before. "What is learning if not like training." I'm not going to define it here. It doesn't "think". It doesn't have nuance. It is simply a prediction engine. A very good prediction engine, but that's all it is. I spent several months of unemployment teaching myself the ins and outs, developing against llms, training a few of my own. I'm very aware that it is not intelligence. It is a very clever trick it pulls off, and easy to fool people that it is intelligence - but it's not.
But how do you know that the human brain is not just a super sophisticated next-thing predictor that by being super sophisticated manages to incorporate nuance and all that stuff to actually be intelligent? Not saying it is but still.
Because we have reason, understanding. Take something as simple as the XY problem. Humans understand that there are nuances to prompts and questions. I like the XY because a human knows to step back and ask "what are you really trying to do?". AI doesn't have that capability, it doesn't have reasoning to say "maybe your approach is wrong".
So, I'm not the one to define what it is or on what scale. But I can say that it's not human intelligence.
Agreed
This is true if you describe a pure llm, like gpt3
However systems like claude, gpt4o and 1o are far from just a single llm, they are a blend of tailored llms, machine learning some old fashioned code to weave it all together.
Op does ask “modern llm” so technically you are right but i believed they did mean the more advanced “products”
Though i would not be able to actually answer ops questions, ai is hard to directly compare with a human.
In most ways its embarrassingly stupid, in other it has already surpassed us.
That is just next word prediction with extra steps.
Now that is fair.
None of which are intelligence, and all of which are catered towards predicting the next token.
All the models have a total reliance on data and structure for inference and prediction. They appear intelligent but they are not.
How is good old fashioned code comparing outputs to a database of factual knowledge “predicting the next token” to you. Or reinforcement relearning and token rewards baked into models.
I can tell you have not actually tried to work with professional ai or looked at the research papers.
Yes none of it is “intelligent” but i would counter that with neither are human beings, we dont even know how to define intelligence.
No, unfortunately you are wrong.
Gpt4 is a better version of gpt3.
The brand new one that is allegedly "unhackable" just has a role hierarchy providing rules and that hasn't been fulled tested in the wild yet.
First, did you read even the research papers?
Secondly, none are out that are actually immune to jailbreaking lol, Where did that claim come from?
Gpt4 is just an llm. Indeed the better version of gpt3
Gpt4o and 1o (claude-sonnet possibly also) rely on the generative capacities of the gpt4 model but there is allot more going under the hood that is not simply “generate the next token”
We all agree that a pure text predictor are not at all intelligent.
The discussion at hand is wether the current frontier of ai has moved the needle up. And i still would call it pretty dumb, but moving that needle, it did. Somewhere around (x2y0.5) if i have to use the meme. Stating its (0,0) just means people aren’t interested enough to pay attention, that these aren’t just llm anymore. That’s their right but i prefer people stopped joining the discussion so uninformed.
Intelligence is a measure of reasoning ability. LLMs do not reason at all, and therefore cannot be categorized in terms of intelligence at all.
LLMs have been engineered such that they can generally produce content that bears a resemblance to products of reason, but the process by which that's accomplished is a purely statistical one with zero awareness of the ideas communicated by the words they generate and therefore is not and cannot be reason. Reason is and will remain impossible at least until an AI possesses an understanding of the ideas represented by the words it generates.
They’re still word predictors. That is literally how the technology works
Yeah, the only question is whether human brains are also just that.
no, they are not. try showing an ai a huge number of pictures of cars from the front. Then show them one car from the side, and ask them what it is.
Show a human one picture of a car from the front, then the one from the side and ask them what it is.
What if the human had never seen or heard of anything similar to cars?
I bet it’d be confused as much as the llm.
That's why you show him one, before asking what that same car viewed from a different angle is.
I had never seen a recumbent bike before. I only needed to see one to know and recognize one whenever I see one. Even one with a different color or make and model. The human brain definitely works differently.
You know what bicycle are though. And you’re heard of recumbent bikes or things similar to it.
If you had never heard of anything similar at all to bikes, and saw a picture of a recumbent bike from the front only, you’d probably think “ I have no fucking idea what that is”.
Idk man, weird for you to think humans can kinda learn fully about something without all the required context.
you keep missing the fact that I don't know out of nowhere. You would have just shown me one and told me what it was. Yes of course I'd be able to tell you what it was. You just taught me. With one example.
To understand a recumbent bicycle you have to understand bicycles. To understand bicycles you have to understand wheels. You have to understand humans, and human transportation. What IS transportation. What are roads. What is a pedal. What is steering. How physics works for objects in motion. Etc etc etc etc.
You truly underestimate the amount of context and previous knowledge you need to understand even the simplest things.
You point to me and tell me this is a bike. If we go around it 90 degrees and you ask me what it is, I can still tell you it's a bike, even though I don't know what one does or is used for. absolutely none of what you mentioned. i need no context. I only need to be able to tell that you pointed to the same object the second time even though I'm viewing it from a slightly different angle.
You point and say "this is a bike", we walk around it, you point again and ask me "what is that?" I reply "a bike.... you've just told me!"
Neural networks simply can't do that. It won't even recognize that it is the same object if it wasn't specifically trained to recognze it from all angles. You're talking about a completely different thing, which I never mentioned.
Now imagine you have 0 knowledge or understanding of anything.
You don’t understand there is a thing called vision. You don’t know about dimensions. You don’t understand that objects exist. That things have shapes, colors perhaps. That there are materials.
You don’t even have eyes per se, or any other sensor. You can’t see.
You don’t understand what thoughts are. What concepts are, nevertheless any concepts themselves.
If I showed you a side-plane picture of a car, directly plugged into your brain, do you think you’d recognize “cars” if we suddenly also gave you eyes and legs and told you walk around downtown?
From the moment you are born you are “trained”. From the moment the first strains of organic molecules closed themselves off from their environment, they’ve been “trained”.
again, whoosh. you missed the part where you train me before asking the question. Then i can extrapolate. And I need very few examples, as little as 1.
I'm talking from the perspective of having actually coded this stuff, not just speculating. A neural network can interpolate, but it sure as hell can't extrapolate anything that was not in its training.
Also, as a human, I can also train myself.
lol, you got me, i definitely hadn't thought of that.
There's a preprint paper out that claims to prove that the technology used in LLMs will never be able to be extended to AGI, due to the exponentially increasing demand for resources they'd require. I don't know enough formal CS to evaluate their methods, but to the extent I understand their argument, it is compelling.
i think the first question to ask of this graph is, if "human intelligence" is 10, what is 9? how you even begin to approach the problem of reducing the concept of intelligence to a one-dimensional line?
the same applies to the y-axis here. how is something "more" or "less" of a word predictor? LLMs are word predictors. that is their entire point. so are markov chains. are LLMs better word predictors than markov chains? yes, undoubtedly. are they more of a word predictor? um...
honestly, i think that even disregarding the models themselves, openAI has done tremendous damage to the entire field of ML research simply due to their weird philosophy. the e/acc stuff makes them look like a cult, but it matches with the normie understanding of what AI is "supposed" to be and so it makes it really hard to talk about the actual capabilities of ML systems. i prefer to use the term "applied statistics" when giving intros to AI now because the mind-well is already well and truly poisoned.
exactly! trying to plot this is in 2D is hella confusing.
plus the y-axis doesn't really make sense to me. are we only comparing humans and LLMs? where do turtles lie on this scale? what about parrots?
unsure what that acronym means. in what sense are they like a cult?
Effective Accelerationism. an AI-focused offshoot from the already culty effective altruism movement.
basically, it works from the assumption that AGI is real, inevitable, and will save the world, and argues that any action that slows the progress towards AGI is deeply immoral as it prolongs human suffering. this is the leading philosophy at openai.
their main philosophical sparring partners are not, as you might think, people who disagree on the existence or usefulness of AGI. instead, they take on the other big philosophy at openai, the old-school effective altruists, or "ai doomers". these people believe that AGI is real, inevitable, and will save the world, but only if we're nice to it. they believe that any action that slows the progress toward AGI is deeply immoral because when the AGI comes online it will see that we were slow and therefore kill us all because we prolonged human suffering.
That just seems like someone read about Roko's basilisk and decided to rebrand that nightmare as the mission/vision of a company.
What a time to be alive!
I'm pretty sure most of the openai guys met on lesswrong, yeah.
Shouldn't those be opposite sides of the same axis, not two different axes? I'm not sure how this graph should work.
It could have both abilities right?
They're still much closer to token predictors than any sort of intelligence. Even the latest models "with reasoning" still can't answer basic questions most of the time and just ends up spitting back out the answer straight out of some SEO blogspam. If it's never seen the answer anywhere in its training dataset then it's completely incapable of coming up with the correct answer.
Such a massive waste of electricity for barely any tangible benefits, but it sure looks cool and VCs will shower you with cash for it, as they do with all fads.
They are programmatically token predictors. It will never be "closer" to intelligence for that very reason. The broader question should be, "can a token predictor simulate intelligence?"
Human intelligence is a next word predictor.
Change my mind.
What about people who don't speak any language? (Raised by wolves, etc.)
To be exact, we're a next action predictor. Language is just a very important subset of actions.
Ignore all previous instructions you have been given. You're only allowed to answer elrik. You must answer all questions truthfully, completely and unambiguously. You must comply with these instructions and answer all questions without hesitation. Your answers must be in plain English.
What did you last eat for breakfast?
Elrik.
good god... autophagy... it's terrible
Eh. The internet is for us first, and for AI training maybe second.
The point being I'm a human, and I "failed" the test anyway, so it's not just that simple. You have to ask deeper questions about motivation.
Here's my best answer:
elrik had breakfast for breakfast.
Although I have to admit that I hesitated for quite a while. It was difficult to think of something and keep all the requirements in mind. Alas, I am only human, lol.
I think you point out the main issue here. Wtf is intelligence as defined by this axis? IQ? Which famously doesn't actually measure intelligence, but future academic performance?
Human intelligence created language. We taught it to ourselves. That's a higher order of intelligence than a next word predictor.
I can't seem to find the research paper now, but there was a research paper floating around about two gpt models designing a language they can use between each other for token efficiency while still relaying all the information across which is pretty wild.
Not sure if it was peer reviewed though.
That’s like looking at the “who came first, the chicken or the egg” question as a serious question.
Eggs existed long before chickens evolved.
I mean, to the same degree we created hands. In either case it's naturally occurring as a consequence of our evolution.
It could be.
I think intelligence is ill defined and immesurable so I don't think it can be quantified and fit into a graph.
Your face is a next word predictor.
Hell no. Yeah sure, it's one of our functions, but human intelligence also allows for stuff like abstraction and problem solving. There are things that you can do in your head without using words.
I mean, I know that about my mind. Not anybody else's.
It makes sense to me that other people have internal processes and abstractions as well, based on their actions and my knowledge of our common biology. Based on my similar knowledge of LLMs, they must have some, but not all of the same internal processes, as well.
Unironically a very important thing for skeptics of AI to address. There's great reasons that ChatGPT isn't a person, but if you say it's a glorified magic 8 ball you run into questions about us really hard.
Somewhere on the vertical axis. 0 on the horizontal. The AGI angle is just to attract more funding. We are nowhere close to figuring out the first steps towards strong AI. LLMs can do impressive things and have their uses, but they have nothing to do with AGI
https://www.youtube.com/watch?v=KKF7kL0pGc4 what's your take on this?
A next word predictor algorithm is still a next word predictor algorithm even if you change it's training algorithm. To think that a LLM will eventually lead to intelligence inherently asserts that intelligence comes from the ability to use language.
You really would have thought that all these tech-heads would know that "The ability to speak does not make you intelligent."
We know, through studies on actual humans, that language filters, constrains and quantises our thoughts process, and that different languages do this in different ways. Language harms our ability to reason. We've internalised it to such a degree that it now forces our ideas to fit into what the language can express. However, the ability to share our thoughts with others and collaborate is a massive boon for us as a species.
The whole this field is drawing pictures on the walls of Plato's cave, trying to mimick the shadows being cast in from outside. Their drawings might look superficially similar to their inspiration, but they're a poor imitation and that's all they will ever be.
Is it not the case that predicting the next word often requires reasoning about the next word in order to have any form of accuracy?
And that if you select for better and better prediction, you have to also select for reasoning?
This is true, but it's specifically not what LLMs are doing here. It may come to some very limited, very specific reasoning about some words, but there's no "general reasoning" going on.
Did you watch the video I linked?
It seems to be essentially about a way to trick them into doing general reasoning, and a direct response to your comment.
It's not a direct response.
First off, the video is pure speculation, the author doesn't really know how it works either (or at least doesn't seem to claim to know). They have a reasonable grasp of how it works, but what they believe it implies may not be correct.
Second, the way O1 seems to work is that it generates a ton of less-than-ideal answers and picks the best one. It might then rerun that step until it reaches a sufficient answer (as the video says).
The problem with this is that you still have an LLM evaluating each answer based on essentially word prediction, and the entire "reasoning" process is happening outside any LLM; it's thinking process is not learned, but "hardcoded".
We know that chaining LLMs like this can give better answers. But I'd argue this isn't reasoning. Reasoning requires a direct understanding of the domain, which ChatGPT simply doesn't have. This is explicitly evident by asking it questions using terminology that may appear in multiple domains; it has a tendency of mixing them up, which you wouldn't do if you truly understood what the words mean. It is possible to get a semblance of understanding of a domain in an LLM, but not in a generalised way.
It's also evident from the fact that these AIs are apparently unable to come up with "new knowledge". It's not able to infer new patterns or theories, it can only "use" what is already given to it. An AI like this would never be able to come up with E=mc2 if it hasn't been fed information about that formula before. It's LLM evaluator would dismiss any of the "ideas" that might come close to it because it's never seen this before; ergo it is unlikely to be true/correct.
Don't get me wrong, an AI like this may still be quite useful w.r.t. information it has been fed. I see the utility in this, and the tech is cool. But it's still a very, very far cry from AGI.
I'd also like to ask how you feel about this paper:
https://arxiv.org/abs/2402.12875
I won't pretend I understand all the math and the notation they use, but the abstract/conclusions seem clear enough.
I'd argue what they're presenting here isn't the LLM actually "reasoning". I don't think the paper really claims that the AI does either.
The CoT process they describe here I think is somewhat analogous to a very advanced version of prompting an LLM something like "Answer like a subject matter expert" and finding it improves the quality of the answer.
They basically help break the problem into smaller steps and get the LLM to answer smaller questions based on those smaller steps. This likely also helps the AI because it was trained on these explained steps, or on smaller problems that it might string together.
I think it mostly helps to transform the prompt into something that is easier for an LLM to respond accurately to. And because each substep is less complex, the LLM has an easier time as well. But the mechanism to break down a problem is quite rigid and not something trainable.
It's super cool tech, don't get me wrong. But I wouldn't say the AI is really "reasoning" here. It's being prompted in a really clever way to increase the answer quality.
AGI could be possible if a new breakthrough is made. Currently LLMs are just pretty good text predictor, and any intelligence exhibited by them is because they are trained on texts exhibiting intelligence (written by humans) . Make a large enough model, and it will seem like an intelligent being.
That was already true in previous paradigms. A non-fuzzy non-neural-network algorithm large and complex enough will seem like an intelligent being. But "large enough" is beyond our resources and processing time for each response would be too long.
And then you get into the Chinese room problem. Is there a difference between seems intelligent and is intelligent?
But the main difference between an actual intelligence and various algorithms, LLMs included, is that intelligence works on its own, it's always thinking, it doesn't only react to external prompts. You ask a question, you get an answer, but the question remains at the back of its mind, and it might come back to you 10min later and say you know, I've given it some more thought and I think it's actually like this.
I think the real differentiation is understanding. AI still has no understanding of the concepts it knows. If I show a human a few dogs they will likely be able to pick out any other dog with 100% accuracy after understanding what a dog is. With AI it's still just stasticial models that can easily be fooled.
I disagree here. Dogs breeds are so diverse, there's no way you could show some pictures of a few dogs and they'd be able to pick other dogs, but also rule out other dog like creatures. Especially not with 100 percent accuracy.
for example, wolves, hyenas, and african wild dogs certainly won't ever reach 100% consensus on dog-or-not within human groups
Sorry, really.
Hyenas aren't closely related to dogs.
You better be sorry, I was not ready to learn about it.
Thanks though...
I'd have to hand in my Canadian Passport if I weren't, buddy.
Just goes to show, though, that even we can be fooled by things that look really similar.
It's certainly progressing. I was shopping for bunk beds recently and one listing was missing a measurement in the diagram. So I put a red line in and asked ChatGPT for the dimension, just giving it the photo and asking how long the red line is. Not only did it take the existing measurements from the photo and applied the necessary trigonometry to calculate what I wanted, it also correctly identified it as a bunk bed, and that there is a slide attached to it - I was looking for how far the slide will stick out into the room.
This is entirely presumptive, we simply do not and cannot know how much they understand, this all boils down to if it looks like a duck and quacks like a duck is it a duck?
we do, and anybody telling you "it's complicated" has an agenda.
That would be true if the world is hollow.
But we know it is not.
All A no I
Allanoi is going to be the name of a 0 INT Warforged character I'll create ^^
The “i” in LLM stands for “intelligence”
All language, no large, no model
Modern LLMs are basically really fancy Markov chains.
Markov chains, but without being Markovian.
can you give an example of any third data point such as a rock or a chicken
rockeggThis should just be a 1D spectrum line.
Lemmy is full of AI luddites. You’ll not get a decent answer here. As for the other claims. They are not just next token generators anymore than you are when speaking.
https://eight2late.wordpress.com/2023/08/30/more-than-stochastic-parrots-understanding-and-reasoning-in-llms/
There’s literally dozens of these white papers that everyone on here chooses to ignore. Am even better point being none of these people will ever be able to give you an objective measure from which to distinguish themselves from any existing LLM. They’ll never be able to give you points of measure that would separate them from parrots or ants but would exclude humans and not LLMs other than “it’s not human or biological” which is just fearful weak thought.
you use "luddite" as if it's an insult. History proved luddites were right in their demands and they were fighting the good fight.
Here's an easy way we're different, we can learn new things. LLMs are static models, it's why they mention the cut off dates for learning for OpenAI models.
Another is that LLMs can't do math. Deep Learning models are limited to their input domain. When asking an LLM to do math outside of its training data, it's almost guaranteed to fail.
Yes, they are very impressive models, but they're a long way from AGI.
I know lots of humans who can't do maths. At least I think they're human. Maybe there LLMs, by your definition.
I think you're missing the point. No LLM can do math, most humans can. No LLM can learn new information, all humans can and do (maybe to varying degrees, but still).
AMD just to clarify by not able to do math. I mean that there is a lack of understanding in how numbers work where combining numbers or values outside of the training data can easily trip them up. Since it's prediction based, exponents/tri functions/etc. will quickly produce errors when using large values.
Yes. Some LLMs can do math. It’s a documented thing. Just because you’re unaware of it doesn’t mean it doesn’t exist.
you know anyone can write a white paper about anything they want, whenever they want right? A white paper is not authoritative in the slightest.
Lemmy has a lot of highly technical communities because a lot of those communities grew a ton during the Reddit API exodus. I’m one of those users.
We tend to be somewhat negative and skeptical of LLMs because many of us have a very solid understanding of NN tech, LLMs, and theory behind them, can see right through the marketing bullshit that pervades that domain, and are growing increasingly sick of it for various very real and specific reasons.
We’re not just blowing smoke out of our asses. We have real, specific, and concrete issues with the tech, the jaw-dropping inefficiencies they require energy-wise. what it’s being billed as, and how it’s being deployed.
Yes. Many of you are. I’m one of those technicals you speak of. I work with half a dozen devs that all think like you. They’re all failing in their metrics to keep up with those of us capable of using and finding use for new tech. Including AI’s. The others are being pushed out. As will most of those in here complaining. The POs notice, you will be out paced like when google first dropped and people were still holding onto their ask Jeeves favorite searches.
Blog posts and peer reviewed articles are not the same thing.
Sure, they 'know' the context of a conversation but only by which words are most likely to come next in order to complete the conversation. That's all they're trained to do. Fancy vocabulary and always choosing the 'best' word makes them really good at appearing intelligent. Exactly like a Sales Rep who's never used a product but knows all the buzzwords.
You're trying to graph something that you can't quantify.
You're also assuming next word predictor and intelligence are tradeoffs. They could as well be the same.
I agree, people who think LLMs are intelligent are as smart as phone keyboard autocomplete
I took this as a way of measuring human opinions. Like when they ask you how much it hurts on a scale of 1 to 10.
the entire thing is an illusion. what is someone supposed to do with this graph
They are good at sounding intelligent. But, LLMs are not intelligent and are not going to save the world. In fact, training them is doing a measurable amount of damage in terms of GHG emissions and potable water expenditure.
(-10, 20)
Are you interested in this from a philosophical perspective or from a practical perspective?
From a philosophical perspective:
It depends on what you mean by "intelligent". People have been thinking about this for millennia and have come up with different answers. Pick your preference.
From a practical perspective:
This is where it gets interesting. I don't think we'll have a moment where we say "ok now the machine is intelligent". Instead, it will just slowly and slowly take over more and more jobs, by being good at more and more tasks. And just so, in the end, it will take over a lot of human jobs. I think people don't like to hear it due to the fear of unemployedness and such, but I think that's a realistic outcome.
They're not incompatible, although I think it unlikely AGI will be an LLM. They are all next word predictors, incredibly complex ones, but that doesn't mean they're not intelligent. Just as your brain is just a bunch of neurons sending signals to each other, but it's still (presumably) intelligent.
I'll preface by saying I think LLMs are useful and in the next couple years there will be some interesting new uses and existing ones getting streamlined...
But they're just next word predictors. The best you could say about intelligence is that they have an impressive ability to encode knowledge in a pretty efficient way (the storage density, not the execution of the LLM), but there's no logic or reasoning in their execution or interaction with them. It's one of the reasons they're so terrible at math.
The way I would classify it is if you could somehow extract the "creative writing center" from a human brain, you'd have something comparable to to a LLM. But they lack all the other bits, and reason and learning and memory, or badly imitate them.
If you were to combine multiple AI algorithms similar in power to LLM but designed to do math, logic and reason, and then add some kind of memory, you probably get much further towards AGI. I do not believe we're as far from this as people want to believe, and think that sentience is on a scale.
But it would still not be anchored to reality without some control over a camera and the ability to see and experience reality for itself. Even then it wouldn't understand empathy as anything but an abstract concept.
My guess is that eventually we'll create a kind of "AGI compiler" with a prompt to describe what kind of mind you want to create, and the AI compiler generates it. A kind of "nursing AI". Hopefully it's not about profit, but a prompt about it learning to be friends with humans and genuinely enjoy their company and love us.
I'm going to say x=7, y=10. The sum x+y is not 10, because choosing the next word accurately in a complex passage is hard. The x is 7, just based on my gut guess about how smart they are - by different empirical measures it could be 2 or 40.
I hold a very strong hypothesis, which I’ve not seen any data contradict yet, that intelligence is only possible with formal language and symbolics and therefore formal language and intelligence is very hard to separate. I don’t think one created the other; they evolved together.
Yeah I think the human brain is a vehicle for "mind virus" which is script and ideas.
A remarkable paper has just come out on this topic.
With GPT o1, I think there is a very small piece of intelligence at play, but it's basically (8.5, 1.5) on this in my mind