Comment on OpenAI Says It's Scanning Users' ChatGPT Conversations and Reporting Content to the Police
Perspectivist@feddit.uk 1 day agoBut it doesn’t understand - at least not in the sense humans do. When you give it a prompt, it breaks it into tokens, matches those against its training data, and generates the most statistically likely continuation. It doesn’t “know” what it’s saying, it’s just producing the next most probable output. That’s why it often fails at simple tasks like counting letters in a word - it isn’t actually reading and analyzing the word, just predicting text. In that sense it’s simulating understanding, not possessing it.
iglou@programming.dev 1 day ago
You’re entering a more philosophical debate than a technical one, because for this point to make any sense, you’d have to define what “understanding” language means for a human in a level as low as what you’re describing for an LLM.
Can you affirm that what a human brain does to understand language is so different to what an LLM does?
I’m not saying an LLM is smart, but saying that it doesn’t understand, when having computers “understand” natural language is the core of NLP, is meh.
Feyd@programming.dev 1 day ago
No they’re not they’re talking purely at a technical level and you’re trying to apply mysticism to it.
iglou@programming.dev 15 hours ago
They are talking at a technical level only on one side of the comparison. It makes the entire discussion pointless. If you’re going to compare the understanding of a neural network and the understanding of a human brain, you have to go into depth on both sides.
Mysticism? Lmao. Where? Do you know what the word means?
Perspectivist@feddit.uk 17 hours ago
You’re right - in the NLP field, LLMs are described as doing “language understanding,” and that’s fine as long as we’re clear what that means. They process natural language input and can generate coherent output, which in a technical sense is a kind of understanding.
But that shouldn’t be confused with human-like understanding. LLMs simulate it statistically, without any grounding in meaning, concepts or reference to the world. That’s why earlier GPT models could produce paragraphs of flawless grammar that, once you read closely, were complete nonsense. They looked like understanding, but nothing underneath was actually tied to reality.
So I’d say both are true: LLMs “understand” in the NLP sense, but it’s not the same thing as human understanding. Mixing those two senses of the word is where people start talking past each other.
iglou@programming.dev 15 hours ago
Of course the “understanding” of an LLM is limited. Because the entire technology is new, and it’s far from being anywhere close to being able to understand to the level of a human.
But I disagree with your understanding of how an LLM works. At its lower level, it’s a bunch on connected artifical neurons, not that different from a human brain. Now please don’t read this as me saying it’s as good as a human brain. It’s definitely not, but its inner workings are not so far. As a matter of fact, there is active effort to make artificial neurons behave as close as possible to a human neuron.
If it was just statistics, it wouldn’t be so difficult to look at the trained model and identify what does what. But just like the human brain, it is incredidbly difficult to understand that. We just have a general idea.
So it does understand, to a limited extent. Just like a human, it won’t understand what it hasn’t been exposed to. And unlike a human, it is exposed to a very limited set of data.
You’re putting the difference between a human’s “understanding” and an LLM’s “understanding” in the meaning of the word “understanding”, which is just a shortcut to say that they can’t be compared. The actual difference is in the scope of understanding.
A lot of the efforts in the AI fields gravitate around imitating a human brain. Which makes sense, as it is the only thing we know that is capable of doing what we want an AI to do. LLMs are no different, but their scope is limited.
Perspectivist@feddit.uk 6 hours ago
I think comparing an LLM to a brain is a category mistake. LLMs aren’t designed to simulate how the brain works - they’re just statistical engines trained on language. Trying to mimic the human brain is a whole different tradition of AI research.
An LLM gives the kind of answers you’d expect from something that understands - but that doesn’t mean it actually does. It acts like something that understands, but that doesn’t mean it does. The danger is sliding from “it acts like” to “it is.” I’m sure it has some kind of world model and is intelligent to an extent, but I think “understands” is too charitable when we’re talking about an LLM.
And about the idea that “if it’s just statistics, we should be able to see how it works” - I think that’s backwards. The only reason it’s so hard to follow is because it’s nothing but raw statistics spread across billions of connections. If it were built on clean, human-readable rules, you could trace them step by step. But with this kind of system, it’s more like staring into noise that just happens to produce meaning when you ask the right question.
I also can’t help laughing a bit at myself for once being the “anti-AI” guy here. Usually I’m the one sticking up for it.
Shanmugha@lemmy.world 13 hours ago
Well, yeah. Humans have these pescy things like concepts, consciousness and thinking above language level. So pesky (sarcasm)
iglou@programming.dev 12 hours ago
That doesn’t answer the question you quoted.
Shanmugha@lemmy.world 12 hours ago
Does it not? Show me how