How do you programmatically check if an LLM is making things up that aren't in the provided search results?

What is a "Golden Dataset" (or Ground Truth set), and how many samples should it ideally contain before you can trust your evaluation metrics?

When would you evaluate a model without having a "correct" answer to compare it against? (e.g., checking for tone or politeness).

How do you measure "Time to First Token" (TTFT) vs. "Total Runtime"? Which one matters more for user experience in a chatbot?