Easy Prompt Engineer questions

34 questions · 5 points each for a correct first answer.

1. 1.Improving a vague promptinstruction-designeasy
2. 2.Few-shot promptingfew-shot-promptingeasy
3. 3.Separating instructions from dataprompt-structureeasy
4. 4.Giving the model a roleinstruction-designeasy
5. 5.Getting a consistent formatoutput-formateasy
6. 6.Zero-shot promptingfew-shot-promptingeasy
7. 7.Reducing made-up answershallucination-reductioneasy
8. 8.Testing a prompt properlyevaluationeasy
9. 9.What a token isllm-fundamentals, tokenseasy
10. 10.What fills the context windowllm-fundamentals, context-windoweasy
11. 11.Raising the temperaturesampling-parameterseasy
12. 12.What top-p controlssampling-parameterseasy
13. 13.Reply cut off mid-sentencellm-fundamentals, api-parameterseasy
14. 14.Stop sequencesapi-parameterseasy
15. 15.The model forgot the last messagellm-fundamentals, conversation-designeasy
16. 16.Questions about last weekllm-fundamentals, rageasy
17. 17.Using XML-style tagsdelimiters, prompt-structureeasy
18. 18.Prefilling the responseprefill, structured-outputeasy
19. 19.Chain-of-thought promptingchain-of-thoughteasy
20. 20.What RAG meansrageasy
21. 21.What embeddings are forembeddings, rageasy
22. 22.How function calling workstool-callingeasy
23. 23.What a hallucination ishallucinationeasy
24. 24.Jailbreak or prompt injectionprompt-injection, llm-safetyeasy
25. 25.Customer data in promptspii-handling, llm-safetyeasy
26. 26.What a golden set isevaluation, golden-setseasy
27. 27.Keeping track of prompt changesprompt-versioningeasy
28. 28.Quote first, then answergrounding, hallucinationeasy
29. 29.Message roles in chat APIsconversation-design, llm-fundamentalseasy
30. 30.Same text, more tokenstokens, multilingual-promptingeasy
31. 31.What you pay forcost-optimization, tokenseasy
32. 32.Temperature 0 still variessampling-parameterseasy
33. 33.Making replies fasterlatency-optimizationeasy
34. 34.Asterisks in text messagesoutput-formatting, instruction-designeasy