6 ms·

Ok, then let us discuss what you define stochastic parrot as.

AndrewDucker 6d ago
Happy to go with the first paragraph from the Wikipedia entry: https://en.wikipedia.org/wiki/Stochastic_parrot https://en.wikipedia.org/wiki/Stochastic_parrot
- hardbass 6d ago
  Suppose a "parrot" factors a product of two 2048 bit primes? Is it still an entity without understanding of the processes considering the search space size?
  - AndrewDucker 6d ago
    Does it have an internal model of what a prime is? If so then it's not parroting. If it doesn't then it is. The whole point is that either they have internal models of the concepts, from experience of those things, or they don't, and are purely working from ungrounded symbols which have no connection to the actual things they're referencing. So, examples of "Look, it did this!" mean nothing to me. What matters is their internal systems.
    - hardbass 6d ago
      What do you think "experience is"? The human brain and body stores experience, like anything else as physical data in itself. So why do you think an LLM can't have experiential knowledge if it is trained to or somehow naturally gains it during training?
      - AndrewDucker 6d ago
        An LLM, as the name implies, only has access to words. It can only model things about those words and the relationships between them. It never has access to the actual things the words represent. So it absolutely has experiential knowledge of words but not of the actual things. (You can absolutely train models on other inputs. I have no objections to those not being stochastic parrots)
        hardbass 5d ago
        Yes but I meant that for instance, how is experience stored in the humans? It is encoded as data in the brain and or rest of body. So I think its possible even a pure textual LLM may have imbibed experiential data somehow during its training. But if your claim is at least that you are fine with thinking AI's that also include physical training are not stochastic parrots, then thats good enough for me.

Happy to go with the first paragraph from the Wikipedia entry: https://en.wikipedia.org/wiki/Stochastic_parrot https://en.wikipedia.org/wiki/Stochastic_parrot

hardbass 6d ago
Suppose a "parrot" factors a product of two 2048 bit primes? Is it still an entity without understanding of the processes considering the search space size?
- AndrewDucker 6d ago
  Does it have an internal model of what a prime is? If so then it's not parroting. If it doesn't then it is. The whole point is that either they have internal models of the concepts, from experience of those things, or they don't, and are purely working from ungrounded symbols which have no connection to the actual things they're referencing. So, examples of "Look, it did this!" mean nothing to me. What matters is their internal systems.
  - hardbass 6d ago
    What do you think "experience is"? The human brain and body stores experience, like anything else as physical data in itself. So why do you think an LLM can't have experiential knowledge if it is trained to or somehow naturally gains it during training?
    - AndrewDucker 6d ago
      An LLM, as the name implies, only has access to words. It can only model things about those words and the relationships between them. It never has access to the actual things the words represent. So it absolutely has experiential knowledge of words but not of the actual things. (You can absolutely train models on other inputs. I have no objections to those not being stochastic parrots)
      - hardbass 5d ago
        Yes but I meant that for instance, how is experience stored in the humans? It is encoded as data in the brain and or rest of body. So I think its possible even a pure textual LLM may have imbibed experiential data somehow during its training. But if your claim is at least that you are fine with thinking AI's that also include physical training are not stochastic parrots, then thats good enough for me.