• @SirGolan
    link
    English
    79 months ago

    Check out this recent paper that finds some evidence that LLMs aren’t just stochastic parrots. They actually develop internal models of things.