Office space meme:

“If y’all could stop calling an LLM “open source” just because they published the weights… that would be great.”

  • Prunebutt@slrpnk.netOP
    link
    fedilink
    arrow-up
    1
    arrow-down
    1
    ·
    2 days ago

    Ok, then my definition givenwas too narrow, when I said “reproducable binaries”. If data claims to be “open source”, then it needs to supply information on how to reproduce it.

    Open data has other criteria, I’m sure.

    • ricecake@sh.itjust.works
      link
      fedilink
      arrow-up
      1
      ·
      1 day ago

      my definition givenwas too narrow

      Yes, that’s what I said when you opted to take the first half of a sentence out of context.

      https://en.wikipedia.org/wiki/Open_data

      The common usage of open data is just that it’s freely shareable.
      Like I said in my initial comment, people frequently use “open source” to refer to it, but it’s such a pervasive error that it hardly worth getting too caught up on and practically doesn’t count as an error anymore.

      Some open data can’t be reproduced by anyone who has access to the data.

      • Prunebutt@slrpnk.netOP
        link
        fedilink
        arrow-up
        1
        arrow-down
        1
        ·
        1 day ago

        I was specifically addressing the use of the phrase “open source”. And the term “open data” doesn’t apply either, since it’s not a dataset that’s distributed, but rather weights of an LLM with data baked into it. That’s neither open source nor open data.