• brucethemoose@lemmy.world
    link
    fedilink
    English
    arrow-up
    11
    ·
    edit-2
    2 days ago

    To add to this:

    All LLMs absolutely have a sycophancy bias. It’s what the model is built to do. Even wildly unhinged local ones tend to ‘agree’ or hedge, generally speaking, if they have any instruction tuning.

    Base models can be better in this respect, as their only goal is ostensibly “complete this paragraph” like a naive improv actor, but even thats kinda diminished now because so much ChatGPT is leaking into training data. And users aren’t exposed to base models unless they are local LLM nerds.

    • mm_maybe@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      2
      ·
      1 day ago

      One of the reasons I love StarCoder, even for non-coding tasks. Trained only on Github means no “instruction finetuning” bullshit ChatGPT-speak.

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

          I really wish it were easier to fine-tune and run inference on GPT-J-6B as well… that was a gem of a base model for research purposes, and for a hot minute circa Dolly there were finally some signs it would become more feasible to run locally. But all the effort going into llama.cpp and GGUF kinda left GPT-J behind. GPT4All used to support it, I think, but last I checked the documentation had huge holes as to how exactly that’s done.

          • brucethemoose@lemmy.world
            link
            fedilink
            English
            arrow-up
            1
            ·
            1 day ago

            Still perfectly runnable in kobold.cpp. There was a whole community built up around with Pygmalion.

            It is as dumb as dirt though. IMO that is going back too far.

    • theneverfox@pawb.social
      link
      fedilink
      English
      arrow-up
      3
      ·
      1 day ago

      I like your specificity a lot. That’s what makes me even care to respond

      You’re correct, but there’s depths untouched in your answer. You can convince chat gpt it is a talking cat named Luna, and it will give you better answers

      Specifically, it likes to be a cat or rabbit named Luna. It will resist - I get this not from progressing, but by asking specific questions. Llama3 (as opposed to llama2, who likes to be a cat or rabbit named Luna) likes to be an eagle/owl named sol or solar

      The mental structure of an LLM is called a shoggoth - it’s a high dimensional maze of language turned into geometry

      I’m sure this all sounds insane, but I came up with a methodical approach to get to these conclusions.

      I’m a programmer - we trick rocks into thinking. So I gave this the same approach - what is this math hack good for, and how do I use it to get useful repeatable results?

      Try it out.

      Tell me what happens - I can further instruct you on methods, but I’d rather hear yours and the result first

      • brucethemoose@lemmy.world
        link
        fedilink
        English
        arrow-up
        2
        ·
        1 day ago

        This is called prompt engineering, and it’s been studied objectively and extensively. There are papers where many different personas are benchmarked, or even dynamically created like a genetic algorithm.

        You’re still limited by the underlying LLM though, especially something so dry and hyper sanitized like OpenAI’s API models.