Hélène Voyer

PromptWise: Studying Literature Using AI

[
[
[

]
]
]

Core Prompt Engineering Techniques

Core Prompt Engineering Techniques

1. Zero-shot Prompting

  • Definition: Asking the AI to perform a task without providing examples.
  • Use case: Quick queries or general tasks.
  • Example:
    “Summarize the key points of the Paris Agreement.”

2. Few-shot Prompting

  • Definition: Providing a few examples to guide the AI’s response.
  • Use case: Tasks requiring specific formatting or tone.
  • Example:
    *“Translate the following phrases into French:
    1. Hello – Bonjour
    2. Thank you – Merci
      Now translate: Good night.”*

3. Chain-of-Thought Prompting

  • Definition: Encouraging the AI to reason step-by-step before answering.
  • Use case: Complex reasoning or math problems.
  • Example:
    “If a train travels 60 km in 1 hour, how far will it travel in 3 hours? Explain your reasoning.”

4. Role-based Prompting

  • Definition: Assigning a role or persona to the AI to shape its behavior.
  • Use case: Simulations, customer service, or expert advice.
  • Example:
    “You are a career coach. Help me prepare for a job interview in tech.”

5. Instruction-based Prompting

  • Definition: Giving clear, direct instructions to guide the AI.
  • Use case: Task automation or structured outputs.
  • Example:
    “List three benefits of remote work in bullet points.”

6. Iterative Prompting

  • Definition: Refining prompts based on previous outputs to improve results.
  • Use case: Creative writing, coding, or troubleshooting.
  • Example:
    “Rewrite this paragraph to be more concise.” → “Now make it more persuasive.”

🧪 Validation Techniques

  • Cross-check facts with trusted sources
  • Ask the same question in different ways to test consistency
  • Use logic checks to ensure reasoning is sound
  • Peer review outputs with others
  • Monitor for bias or hallucinations

🛠️ Tools & Resources

Tools

  • OpenAI Playground – Test and refine prompts
  • PromptHero – Prompt templates and examples
  • FlowGPT – Community-curated prompt library
  • LangChain / LlamaIndex – Build prompt-driven apps
  • Notebooks / Code Interpreter – Validate and test outputs

Learning Resources