• eatCasserole@lemmy.worldM
    link
    fedilink
    arrow-up
    4
    ·
    10 months ago

    “factual accuracy in large language models remains an area of active research”

    This really says a lot.

  • alyth@lemmy.world
    link
    fedilink
    arrow-up
    4
    arrow-down
    2
    ·
    10 months ago

    Are they gonna file a complaint against Meta next because the Llama weights encode inaccurate facts on a person? File a complaint against the inference libraries too?