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.
A remarkable paper has just come out on this topic.
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.
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.
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.
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.
Your face is a next word predictor.
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.
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?
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.
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.
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?
That would be true if the world is hollow.
But we know it is not.
we do, and anybody telling you “it’s complicated” has an agenda.
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.
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.
Wondering if Modern LLMs like GPT4, Claude Sonnet and llama 3 are closer to human intelligence or next word predictor.
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.
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.
the entire thing is an illusion. what is someone supposed to do with this graph
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 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.
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.
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.
