Your question is Testing Non-Deterministic LLMs. 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 do you write automated tests to evaluate the performance, reliability, and accuracy of non-deterministic LLM outputs?