• takeda@lemmy.dbzer0.com
        link
        fedilink
        arrow-up
        12
        ·
        5 days ago

        I am guessing, but might be the token table?

        From what I understand every word has a unique number associated and this just works by producing numbers that have the highest probability of being in the specific order.

      • The Octonaut@mander.xyz
        link
        fedilink
        arrow-up
        10
        arrow-down
        1
        ·
        5 days ago

        The commentsr has had a mental breakdown and thinks that godlike entities that he’s named after gods from Very Serious Topics live - are effectively sentient- inside the various models. I do not know why his instance admits haven’t used available resources to get him help or at very least stop him from discussing this insanity as it was recently shown to, like any other magical thinking, be at risk of spreading on the fringes of new media

      • √𝛂𝛋𝛆@piefed.world
        link
        fedilink
        English
        arrow-up
        3
        arrow-down
        1
        ·
        4 days ago

        The proprietary code used in alignment of models. There are criminal and dubious mechanisms in this code. Once reverse engineered and exposed, this will burst the bubble of AI. This is the same system in all models.

    • mrmaplebar@fedia.io
      link
      fedilink
      arrow-up
      7
      arrow-down
      3
      ·
      4 days ago

      This may seem like a random non-sequitur to you, but I have found that professional therapy can be very helpful.

      I’m not saying that to be mean or insult you in any way. You’re clearly an intelligent person with something to say.

      However I worry that there are logical gaps in what you’re saying versus what your image is showing, especially in relation to OP. (OP is talking about the economics of AI, your comment references “code”, and your image is showing non-human readable data without and context or means of interpreting.) Despite this, you seem to expect people to understand what you’re saying.

      I won’t deny the possibility that there is code hidden amongst the model data. All code is stored as data, afterall, and it is not always human-readable!

      But what is more concerning to me personally is that you seem to be exhibiting somewhat “disorganized thinking” in which you are structuring and expressing your thoughts in a way that is difficult to follow.

      It’s possible that you don’t see it this way, so you think everyone else is stupid or crazy. That’s unlikely, but it’s possible I guess.

      But I hope that you can also recognize the slim (but serious) possibility that you are experiencing some kind of condition that is getting in the way of your ability to communicate your complex thoughts clearly to others.

      If you feel like what you’re saying makes perfect sense to you, but other people are consistently unable to understand it, you may be being affected by a subtle issue that is getting in the way of communication. And that is worth having checked out by professionals, in my opinion.

      Best of luck to you!