How do you calculate the ROI of a prompt change? If a new prompt is 5% more accurate but 50% more expensive in tokens, how do you decide if it’s worth it?

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

Why is a standard unit test (asserting that output == "expected") often a bad way to test an LLM? How do you handle a model that gives three different, but correct, answers to the same prompt?

If your retriever returns 5 documents but only 1 was actually related to answering the question, how do you penalize the retriever for the "noise"?