6 ms·
You're confusing the training method with the internal process. If I had you repeatedly attempt to learn how to make believable completions of partial document
by aetherson 3mo ago
You're confusing the training method with the internal process. If I had you repeatedly attempt to learn how to make believable completions of partial documents about a given topic, you would eventually learn things about that topic and could use your knowledge to create more believable completions of documents about that topic.
- operatingthetan 3mo agoLLMs do not learn. You put it out to pasture and create a new one. "Memory" in a session is essentially a context window party trick.
- aspenmartin 3mo agoThey do learn in context, and very sample efficiently. Continual learning is active area of research and we sort of already have something resembling it with persistent context. So yes they do learn.
- operatingthetan 3mo agoI consider that to be the illusion of learning. You are not wrong, I think they may actually learn in the future though. But not today.
- aspenmartin 3mo agoThat’s strange to me, what would you define as learning?
- FromTheFirstIn 3mo agoTo acquire new knowledge and build your understanding. They don’t understand so they can’t learn
- operatingthetan 3mo agoThank you for saying succinctly what I could not. If your consciousness and knowledge fundamentally does not change from your ongoing experience, then you are not learning. This is how the LLM currently functions.
- aspenmartin 3mo agoYou’re describing the problem of continual learning. As I said their “consciousness” for lack of a better term and knowledge does already change from ongoing experience in context which is another of saying for only a short window, today. They are ephemeral, sort of, but that’s a temporary limitation.
- FromTheFirstIn 3mo agoI think if your definition of consciousness can fit these things then you’re more open minded than I care to be. Consciousness isn’t really guessing the next thing to say- it’s hard to say what it is, obviously, but blindly feeling forwards with each new conversation doesn’t seem like consciousness or learning to me.
- aspenmartin 3mo agoWe aren't talking about consciousness, we're talking about learning. > Consciousness isn’t really guessing the next thing to say- I don't know what consciousness is either and these debates are a dumpster fire when they happen, but it sounds like you're pulling forward this "LLMs are just predicting the next token" (true by construction) implies that they can't learn or reason or be conscious (2/3 are wrong, the last one isn't falsifiable without a useful definition).
- FromTheFirstIn 3mo agoYes, I think we simply disagree. I think you know what LLMs are based on our other thread and if you think something important happens when you get a large enough context vector I don’t think there’s much I can say to change your mind. It seems unlikely to me!
- lemiffe 3mo agoThe LLM itself doesn't, but agents can research, compare, add to their memory, and use that to narrow the results down to a probabilistically higher set of outputs; I have used an LLM for my own MRI results and it was nearly spot-on, verified by a subsequent visit to a specialist. YMMV as they say. But I do believe we are entering the era where LLMs are considering past interactions and long context windows to inform it of personal preferences and history in order to output more accurate results.
- chiply314 3mo agoThey already learned. A lot or basically everything evern written and available digital. And context window work very well. You can 'teach' an llm a new programming lanuage and other things through it.
- aetherson 3mo agoThey learn during training, which is what we're talking about.
- operatingthetan 3mo ago>which is what we're talking about. You are anyway, I don't see anyone up the chain saying that.
- goodpoint 3mo agobelievable != true
- operatingthetan 3mo agoA very important callout. It's the crux of the whole thing really. Humans are easily susceptible to deception by statements that are structured to be believable.
- aetherson 3mo agoSure. But that's not the subject.
- operatingthetan 3mo agoPlease stop trying to police what the subject is to suit your own arguments.
- fhdkweig 3mo agoThis is what Stephen Colbert called "truthiness". People want to believe what they feel is true even if it is directly contradicted by evidence. https://en.wikipedia.org/wiki/Truthiness https://en.wikipedia.org/wiki/Truthiness