1. What does 'few-shot prompting' mean?
2. What is the main benefit of chain-of-thought (CoT) prompting?
3. Which prompting technique asks the model to imagine it is an expert in a domain?
4. What is prompt injection?
5. What is 'zero-shot chain-of-thought' prompting?
6. Why should output format instructions be placed near the end of a long system prompt?
7. What is the purpose of a 'delimiter' in a structured prompt?
8. What does 'self-consistency' prompting do?