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?

Instead of checking for exact words, how would you use BERTScore or Cosine Similarity of embeddings to evaluate if an LLM's summary is accurate?

Guardrails add an extra check. How do you evaluate if the safety benefit of a guardrail outweighs the 200ms latency penalty it adds?

How would you automate the process of trying to make your model "break" or "hallucinate"?