Trying to build a conversational AI chat assistant. Most of the models like Vicuna and LLaMa 13B are just inundating me with a bunch a text.

I’m trying to emulate a conversation like you’d have with your friend. Not just be bombarded with a bunch of text to read and information to process.

Which models are good at this that can be fine-tuned for a specific usecase?

  • SomeOddCodeGuy@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    I have yet to see anything dethrone Mythomax for the 13b range for conversations. Lots of things become the new shiny, then they fade away… but Mythomax continues to live on.

    • LocoLanguageModel@alien.topB
      link
      fedilink
      English
      arrow-up
      1
      ·
      1 year ago

      Mythomax is great but super slutty.

      Sometimes I like the challenge of getting an LLM to be inappropriate, and Mythomax just jumps right in!

  • vatsadev@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    Open source -> Mistral instruct worked great for me, Zephyr alpha was crazy aligned, while beta was better

    Closed Source -> Inflections Pi is smooth! Pray for API access