• jacksilver
      link
      fedilink
      English
      arrow-up
      2
      ·
      19 days ago

      Its believed (probably more just fact) that a lot of the open source models leverage larger models. Either directly through distiallation (meaning training a model directly off of another) or by using the larger models for knowledge distiallation (basically using the larger models as “ground truth” for training workflows).

      Either way, this means the open source models would always be “a few months behind” as they rely on the progress of the closed source models to improve themselves.

      • venusaur
        link
        fedilink
        English
        arrow-up
        1
        ·
        19 days ago

        I see. Is there some way to protect models from being trained on them aside from just removing it from the market?

        • jacksilver
          link
          fedilink
          English
          arrow-up
          2
          ·
          19 days ago

          From what I’ve read, it seems like a lot of the companies monitor API calls and will posion responses for suspicious traffic. This basically means they’ll modify the output to be harmful to the scrapper.

          Outside of that they can block accounts/IPs, limit what is returned (just the final answer vs all reasoning steps, etc.). However it’s a bit like all cyber security and an evolving game of cat and mouse.

          • venusaur
            link
            fedilink
            English
            arrow-up
            1
            ·
            18 days ago

            How was Qwen previously able to train on Claude? Prompt injection?

            • jacksilver
              link
              fedilink
              English
              arrow-up
              1
              ·
              18 days ago

              I’m not caught up on all the details, but as I said this has been an ongoing game of cat and mouse with people looking to leverage other models to improve their own (this isn’t just a qwen vs Claude thing and everyone if pointing fingers at everyone). If I had to guess it’s only been more recently that AI companies are actively fighting these issues, but even then I imagine they’re not 100% effective.