[cube00]

"Large language models don’t behave like people, even though we may expect them to"

https://news.mit.edu/2024/large-language-models-dont-behave-...