Meaning without reference in large language models (opens in new tab)

Lots of excellent references showing LLMs have internal models. Good precis of the philosophical argument that meaning doesn't require reference.