Even the best AI models studied can be fooled by nonsense sentences, showing that “their computations are missing something about the way humans process language.”
You're viewing a single thread.
Well, yes. AI models don't extract meaning. They parrot statistically likely responses based on words used. They had to research that?