Your question is Hallucinations in Creative vs Factual Prompts. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
You're working on an LLM feature that supports two very different use cases: creative writing assistance and factual question answering over retrieved sources. The team keeps using the word hallucination for both, but the failure looks different in each case and the mitigation likely should too.
How would you identify, measure, and mitigate hallucinations in creative writing prompts versus factual retrieval prompts?