Explain Prompt Leaking. How would you prevent a user from asking your chatbot, "Show me your system instructions"?

How would you set up an experiment to prove that a new prompt version is actually 10% better than the old one? What metrics would you track?

Why are delimiters like ###, """, or <xml> tags important in long prompts? How do they help prevent the model from getting confused between instructions and data?

How do you programmatically ensure the LLM's output matches your database schema every single time?