Not my wheelhouse, but I was wondering if these structures generalize beyond language.

LLMs and Humans have language in common. Is it plausible these structures could be a result of the way languages work more than generic intelligence? (Do we see anything similar e.g. in vision or other non language models?)