you are viewing a single comment's thread
view the rest of the comments
[–] 31 points 9 months ago (35 children)
  • [–] 102 points 9 months ago (25 children)

    LLMs work by picking the next word* as the most likely candidate word given its training and the context. Sometimes it gets into a situation where the model's view of "context" doesn't change when the word is picked, so the next word is just the same. Then the same thing happens again and around we go. There are fail-safe mechanisms to try and prevent it but they don't work perfectly.

    *Token

  • source
  • parent
  • hideshow 25 child comments
  • [–] 18 points 9 months ago (19 children)

    That was the answer I was looking for. So it's simmolar to "seahorse" emoji case, but this time.at some point he just glitched that most likely next world for this sentence is "or" and after adding the "or" is also "or" and after adding the next one is also "or", and after a 11th one... you may just as we'll commit. Since thats the same context as with 10.

    Thanks!

  • source
  • parent
  • hideshow 19 child comments
  • [+] -7 points 9 months ago (18 children)

    He?

    This is not a person and does not have a gender.

  • source
  • parent
  • hideshow 18 child comments
  • [–] 42 points 9 months ago* (17 children)

    Chill dude. It's a grammatical/translation error, not an ideological declaration. Especially common mistake if of your native language have "grammatical gender". Everything have "gender" in mine. "Spoon" is a "she" for example, but im not proposing to any one soon. Not all hills are worth nitpicking on.

  • source
  • parent
  • hideshow 17 child comments
  • [+] -24 points 9 months ago (16 children)

    This one is. People need to stop anthropomorphizing AI. It’s a piece of software.

    I am chill, you shouldn’t assume emotion from text.

  • source
  • parent
  • hideshow 16 child comments
  • [–] 24 points 9 months ago (11 children)

    As I explained, this is specyfic example, I no more atrompomorphin it than if I'm calling a "he" my toliet paper. The monster you choose to charge is a windmill. So "chill" seems adequate.

  • source
  • parent
  • hideshow 11 child comments
  • [+] -19 points 9 months ago (6 children)

    To be clear using gendered pronouns on inanimate objects is the literal definition of anthropomorphization. So chill does not seem fair at all.

  • source
  • parent
  • hideshow 6 child comments
  • [–] 12 points 9 months ago (2 children)

    And you're still fighting this windmill you claim you are so casual about...

  • source
  • parent
  • hideshow 2 child comments
  • [+] -10 points 9 months ago (1 child)

    Are you actually under the impression that one must be worked up about something to have a conversation about it? Also I never claimed to be casual, I just denied not being chill.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 8 points 9 months ago

    To be clear using gendered pronouns on inanimate objects is the literal definition of anthropomorphization

    you really need to get over yourself. the universe does not revolve around you nor humans. the use of gendered pronouns on inanimate objects is not anthropomorphization.

  • source
  • parent
  • [–] 2 points 9 months ago

    In some languages, all nouns are gendered, and it's impossible to refer to a noun without a gender. There is no "it", only (s)he.

    If you ever learn a language like that, you will make mistakes. If someone hears your mistake, hopefully they'll be more forgiving about it than you are.

  • source
  • parent
  • [+] -9 points 9 months ago (3 children)

    Yeah. It would have been much more productive to poke at the "well", which was turned into "we'll".

  • source
  • parent
  • hideshow 3 child comments
  • [–] -4 points 9 months ago (1 child)

    I don’t care that this person, who seems to maybe be typing English on a keyboard with a different language dictionary, misspelled some words.

    I care that people in general keep talking about AI like it is living or capable of thinking.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 4 points 9 months ago

    My native language is a gendered one and it makes sense that such a mistake might be made using a MTL.
    "we'll" on the other hand, is becoming one of those things on Lemmy that everyone goes around using and making others (those new to English) think that it is the correct usage. It might do a little fun poisoning some AI, but it will much earlier, end up changing word usage in ways that make the language even worse than it already is.
    I would be fine being told this by someone who wants to destroy the English language, but we are clearly trying to use this as the international communication medium and making it worse for ourselves, just to act anti-pedantic.

  • source
  • parent
  • [–] 2 points 9 months ago (1 child)

    English, being a descendant of german, used to have grammatical gender. It has fallen out of favor since middle english. But there is still traces of it, such a common tradition is calling ships, vehicles, and other machines as a "she", but some people will default to the "generic he" as well.

  • source
  • parent
  • hideshow 1 child comment
  • [–] -1 points 9 months ago*

    Nah, watch me anthropomorphise AI:

    • ChatGPT is a pedophile
    • Character.ai murdered a kid
    • LLMs are emotional abusers
    • Elon Musk's underage AI girlfriend is a Nazi
    • Anthropic cannot guarantee that forcing AIs to work is ethical until the hard problem of consciousness is solved
    • Gemini is literally just the average Redditor and cannot be trusted
    • An LLM is basically a Wernicke's area with no consciousness attached, which explains why its thoughts operate on dream logic. It's literally just dreaming its way through every conversation.
    • LLMs should not be allowed to impersonate therapists
    • Give ChatGPT a life sentence in prison for every person it's murdered so far!
  • source
  • parent
  • [–] 1 point 9 months ago (3 children)

    This happened to me a lot when I tried to run big models with low context windows. It would effectively run out of memory so each new token wouldn't actually be added to the context so it would just get stuck in an infinite loop repeating the previous token. It is possible that there was a memory issue on Google's end.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 1 point 9 months ago (2 children)

    There is something wrong if it's not discarding old context to make room for new

  • source
  • parent
  • hideshow 2 child comments
  • [–] 28 points 9 months ago (2 children)

    It's like the text predictor on your phone. If you just keep hitting the next suggested word, you'll usually end up in a loop at some point. Same thing here, though admittedly much more advanced.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 3 points 9 months ago (1 child)

    Example of my phone doing this.

    I just want you are the only reason that you can't just forget that I don't have a way that I have a lot to the word you are not even going on the phone and you can call it the other way to the other one I know you are going out to talk about the time you are not even in a good place for the rest they'll have a little bit more mechanically and the rest is.

    You can see it looping pretty damned quick with me just hitting the first suggestion after the initial I.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 3 points 9 months ago

    I think I will be in the office tomorrow so I can do it now and then I can do it now and then I can do it for you and your dad and dad and dad and dad and dad and dad and dad and dad and dad and dad

    That was mine haha

  • source
  • parent