• 0 Posts
  • 29 Comments
Joined 2 years ago
cake
Cake day: August 2nd, 2023

help-circle





  • lily33@lemm.eetoTechnology@lemmy.world*Permanently Deleted*
    link
    fedilink
    English
    arrow-up
    4
    ·
    edit-2
    4 months ago

    Any accessibility service will also see the “hidden links”, and while a blind person with a screen reader will notice if they wonder off into generated pages, it will waste their time too. Especially if they don’t know about such “feature” they’ll be very confused.

    Also, I don’t know about you, but I absolutely have a use for crawling X, Google maps, Reddit, YouTube, and getting information from there without interacting with the service myself.








  • lily33@lemm.eetoOpen Source@lemmy.mlProton's biased article on Deepseek
    link
    fedilink
    arrow-up
    4
    arrow-down
    1
    ·
    edit-2
    6 months ago

    What makes these consumer-oriented models different is that that rather than being trained on raw data, they are trained on synthetic data from pre-existing models. That’s what the “Qwen” or “Llama” parts mean in the name. The 7B model is trained on synthetic data produced by Qwen, so it is effectively a compressed version of Qen. However, neither Qwen nor Llama can “reason,” they do not have an internal monologue.

    You got that backwards. They’re other models - qwen or llama - fine-tuned on synthetic data generated by Deepseek-R1. Specifically, reasoning data, so that they can learn some of its reasoning ability.

    But the base model - and so the base capability there - is that of the corresponding qwen or llama model. Calling them “Deepseek-R1-something” doesn’t change what they fundamentally are, it’s just marketing.