Maybe I am far from the AI research space, but conflating the statistical processes going on in a LLM with terms like ‘mind’, ‘chain-of-thought’, ‘internal neural activations’, and even worse “… is saying the word, just that the word is on it’s mind” makes me feel that this is a buzzword hype-piece rather than an actual paper
There is a very strong effort to anthropomorphise the model in this abstract, which should be a very big no-no for any selfrespecting unbiased researcher.
Maybe I am far from the AI research space, but conflating the statistical processes going on in a LLM with terms like ‘mind’, ‘chain-of-thought’, ‘internal neural activations’, and even worse “… is saying the word, just that the word is on it’s mind” makes me feel that this is a buzzword hype-piece rather than an actual paper
There is a very strong effort to anthropomorphise the model in this abstract, which should be a very big no-no for any selfrespecting unbiased researcher.