Your question is LLM-as-a-Judge Evaluation. 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).
How would you evaluate an interactive, non-deterministic generative AI experience using LLM-as-a-Judge and automated benchmarking?