LLMs

Event knowledge in language models: the gap between the impossible and the unlikely