Learn how to use feedback prompts to help LLMs self-correct their own errors - when it works, when it fails, and how to implement it without falling into overconfidence traps.