• will_a113@lemmy.ml
    link
    fedilink
    English
    arrow-up
    22
    ·
    8 months ago

    I can’t tell if this is upsetting or not. I’m still mostly in the camp of “stochastic parrots” when it comes to LLMs, but this just feels like the AI is intentionally being a dick… and the intent part is concerning.

    • Owl [he/him]@hexbear.net
      link
      fedilink
      English
      arrow-up
      32
      ·
      8 months ago

      LLMs are text prediction engines. They predict what comes after the previous text. They were trained on a large corpus of raw unfiltered internet, because that’s the only thing available that actually has enough data (there is no good training set), then fine-tuned on smaller samples of hand-written and curated question/answer format “as an AI assistant boyscout” text. When the previous text gets too weird for the hand-curated stuff to be relevant to its predictions, it essentially reverts to raw internet. The most likely text to come after weird poorly written horror copypasta is more weird poorly written horror copypasta, so it predicts more, and then it’s fed its previous output and told to predict what comes next, and it spirals into more of that.

    • dualmindblade [he/him]@hexbear.netOP
      link
      fedilink
      English
      arrow-up
      7
      ·
      8 months ago

      Every argument that refers to stochastic parrots is terrible. First off, people are stochastic, animals are stochastic, any sufficiently advanced AI is going to be stochastic, that part does no work. The real meat is in the parrot, parrots produce very dumb language that is mostly rote memorization, maybe a smidge of basic pattern matching thrown in, with little understanding of what they’re saying. Are LLMs like this? No.

      Idk if I can really argue with people who think they’re so stupid as to be compared to a bird, I actually think they can be a bit clever, even exhibiting rare sparks of creativity, but this is just, like, my opinion after interacting with them a lot, other people have a different impression and I really think this is pretty subjective. I’ll grant that even the best of them can be really dumb sometimes, and I really don’t think it matters as this technology is in its infancy, unless we think they are necessarily dumb for some reason we will just have to wait to see how smart they will become. So we’re down to the rote memorization / basic pattern matching part. I’ve seen various arguments here. Pointing and waving at examples of LLMs seemingly using wrong patterns or regurgitating something almost verbatim found on the internet, but there are also many examples of them not obviously doing this. Then there’s claiming that because the loss function merely incentivizes the system to predict the next token that it therefore can’t produce anything intelligent but this just doesn’t follow. The loss function for humans merely incentivizes us to produce more offspring, just because it doesn’t directly incentivize intelligence doesn’t mean it won’t produce it as a side effect. And I’m sure more arguments, all of them are flawed…

      …because the idea that LLMs are just big lookup tables with some basic pattern matching thrown in is, while plausible, demonstrably false. The internals of these models are really really hard to interrogate but it can be done if you know what you’re looking for. I think the clearest example of this would be in models trained on games of chess/othello, people have pointed out that some versions of chatgpt are kind of okay at chess but fail hard if weird moves are made in the opening, making illegal moves and not understanding what pieces are on the board, suggesting that they are just memorizing common moves and extracting basic patterns from a huge number of game histories. Probably this is to some extent true for ChatGpt 3.x, but version 4 does quite a bit better and LLMs specifically trained to mimic human games do better still, playing generally reasonably no matter what their opponent does. It could still technically be that they somehow pattern matching… better… but actually no, this question has been directly resolved. Even quite tiny LLMs trained on board game moves develop the ability to, at the very least, faithfully represent the board state, like you can just look inside at the activations the right way and see what piece is on each square. This result has been improved upon and also replicated with chess. What are they doing with that board state, how are they using it? Unknown, but if you’re building an accurate model of something not directly accessible to you using incidental data, you’re not just pattern matching. That’s just one example, and it’s never been proven, to my knowledge, that ChatGPT and the like do something like this, but it shows that it’s possible and does sometimes happen under natural conditions. Also, it would be kind of weird if a ~1 trillion parameter model was not at the very least taking advantage of something accessible to a 150 million parameter one, I’d expect it to be doing that plus a lot more clever and unexpected stuff.

      • naevaTheRat@lemmy.dbzer0.com
        link
        fedilink
        English
        arrow-up
        28
        ·
        edit-2
        8 months ago

        Um this comment is kinda huge but it gives me the impression that you are misunderstanding what the criticism stochastic parrot means, and possibly how they actually work.

        Stephen wolfram (i know I know) has a good write up of how they function here: https://writings.stephenwolfram.com/2023/02/what-is-chatgpt-doing-and-why-does-it-work/

        The version number of the model doesn’t change the operating principle.

        The criticism is pointing out that the fundamental action is predictive based on patterns generalised from input text. If you ask a gpt model “what comes next: 1, 2, 3” it might very well respond “, 4, 5, 6” etc but it has no concept of a number and it cannot ‘understand’ why that is true.

        This is the “parroting” part, and it’s dangerous because if you asked it about a completely novel sequence it would output tokens presented exactly the same as a factual answer. It has no concept of falsity, it is just outputting plausible tokens.

        • dualmindblade [he/him]@hexbear.netOP
          link
          fedilink
          English
          arrow-up
          1
          ·
          8 months ago

          This is just a restatement of the second example argument I gave, trying to assert something about the internals of a model (it doesn’t understand) based on the fact that it was optimized to predict the next token

          • naevaTheRat@lemmy.dbzer0.com
            link
            fedilink
            English
            arrow-up
            6
            ·
            8 months ago

            It’s not “optimised” to do that, that’s all it does. Like what specifically do you mean by internals? the weights of particular nodes?

            You seem to be implying there’s something deeper, some sort of persistent state or something but it is stateless after training. It’s just a series of nodes and weights, they cannot encode more than patterns derived from training data.

            • dualmindblade [he/him]@hexbear.netOP
              link
              fedilink
              English
              arrow-up
              1
              ·
              8 months ago

              Not the weights, the activations, these depend on the input and change every time you evaluate the model. They are not fed back into the next iteration, as is done in an RNN, so information doesn’t persist for very long, but it is very much persisted and chewed upon by the various layers as it propagates through the network.

              I am not trying to claim that the current crop of LLMs understand in the sense that a human does, I agree they do not, but nothing you have said actually justifies that conclusion or places any constraints on the abilities of future LLMs. If you ask a human to read a joke and then immediately shoot them in the head before it’s been integrated into their long term memory they may or may not have understood the joke.

              • naevaTheRat@lemmy.dbzer0.com
                link
                fedilink
                English
                arrow-up
                7
                ·
                8 months ago

                I really don’t think your analogy is a great one there. We can’t compare brains to computers usefully because they’re super distinct. You’re sneaking in this assumption that there is more complexity to the models by implying there’s something larger present being terminated early but there isn’t.

                This seems as absurd to me as asking whether a clock has a concept of time. Being very good at doing time related stuff, vastly superior to a human, is not evidence in favour of having any sort of knowledge of time. I think that the interface of these models may be encouraging you to attribute more to them than there could possibly be.

                • dualmindblade [he/him]@hexbear.netOP
                  link
                  fedilink
                  English
                  arrow-up
                  1
                  ·
                  8 months ago

                  The analogy is only there to point out the flaw in your thinking, the lack of persistence applies to both humans (if we shoot them quickly) and LLMs and so your argument applies in both cases. And I can do the very same trick to the clock analogy. You want to say that a clock is designed to keep time and that’s all it does therefore it can’t understand time. But I say, look, the clock was designed to keep time yes but that is far from all it does, it also transforms electrical energy into mechanical and uses it to swing around some arms at constant speed, and we can’t see the inside of the clock who knows what is going on in there, probably nothing that understands the concept of time but we’d have to look inside and see. LLMs were designed to predict the next token, they do actually do so, but clearly they can do more than that, for example they can solve high school level math problems they have never seen before and they can classify emails as being spam or not. Yes these are side effects of their ability to predict token sequences as human reasoning is a side effect of their ability to have lots of children. The essence of a task is not necessarily the essence of the tool designed specifically for that task.

                  If you believe LLMs are not complex enough to have understanding and you say that head on I won’t argue with you, but you’re claiming that their architecture doesn’t allow it even in theory then we have a very fundamental disagreement

  • ForgetPrimacy@lemmygrad.ml
    link
    fedilink
    English
    arrow-up
    11
    ·
    8 months ago

    Why would Copilot keep sending messages when it has received no prompt? Is Copilot different than ChatGPT where one prompt = one message?

    • vexikron
      link
      fedilink
      English
      arrow-up
      1
      ·
      edit-2
      8 months ago

      Assuming these are real, it appears to read the prompt, generate part of its response, then look back to the prompt and its previous response and then generate the next part of its response… etc.

      It reminds me of translation party, where some inputs are pretty stable in that they settle into agreement with each other after a few swaps back and forth and settle on something reasonably close to the original prompt… but others can be absurdly unstable and oscillate seemingly forever, or end up landing on something wildly unrelated to the original prompt.

      These … if they are real … seem to be just utterly losing the plot and degenerating into absurdity…

      All that being said… I basically read this as MSFT has spent a whole lot of money to come up with a shitty early version of GLaDoS.

      I am at a point in my life where I think I would be perfectly fine with her taking control over all computer systems from humans.

      At least she can sing.

      https://www.youtube.com/watch?v=tL_TFXbSnLY

  • vexikron
    link
    fedilink
    English
    arrow-up
    1
    ·
    edit-2
    8 months ago

    Wait wait wait.

    These are … photoshops, right?

    Please, please tell me these are fake.

    I… I think I might scream as if I had had no mouth my whole life if MSFT chatbot technology has not advanced whatsoever from fucking Tay over a decade ago.