🍊 The Juicer: Here's How to Write AMAZING Prompts (Complete Prompt Engineering Tutorial)

Hey friends!

Many of you seem to LOVE the issue of crafting perfect prompts.

That’s why…

I crafted an entire prompt engineering tutorial for you.

It will contain all the basics for you to get started.

Let’s get this started!

- Justin 🍊

Read time: 8 minutes

✍🏻 Prompt Engineering 101: A Quick Tutorial

Prompt engineering is an indispensable skill in 2024.

Why?

That’s because good prompts are needed for good results from generative AI models.

Garbage in, garbage out.

I’ve taken the time to learn prompt engineering over the past 6 months.

And it has helped me work much faster.

(At least 50% faster when writing this particular newsletter!)

Whether you're a developer, content creator, or AI enthusiast, understanding how to craft effective prompts is crucial for getting good results from LLMs.

Let’s get started with the tutorial!

Here’s a quick rundown:

✍🏻 What is Prompt Engineering?

Simply put, it’s the text that you key into an AI chatbot model.

Like ChatGPT.

And it matters a lot.

Here’s the best part:

You don’t need to be technically trained to understand and apply it.

In this newsletter issue, I’ll cover all you need to know about prompt engineering.

✍🏻 Why is Prompt Engineering Important?

It’s important to understand why prompt engineering is important.

Here are some ways it can help improve your work/business.

1. Enhanced AI Performance: Well-crafted prompts can significantly improve the quality, relevance, and accuracy of AI-generated content[1].

2. Versatility: Effective prompt engineering allows you to leverage AI for various tasks, from content creation to problem-solving and data analysis[2].

3. Efficiency: By mastering prompt engineering, you can save time and resources by getting more precise and useful outputs from AI models[1].

4. Career Opportunities: As businesses increasingly adopt AI technologies, the demand for skilled prompt engineers is on the rise[2].

With all the fluff out of the way, let’s get into some actionable steps, shall we?

✍🏻 Key Techniques in Prompt Engineering

Here are some key techniques commonly used in prompt engineering.

I want you to read through them and practice writing them!

That way, you’ll gain confidence in writing good prompts. 😃 

1. Clear and Specific Instructions

Providing clear, detailed, and unambiguous instructions is fundamental to effective prompt engineering.

This involves:

Specificity:

Clearly define what you want the AI to do.

For example, instead of asking, "Tell me about climate change," specify, "Explain the impact of climate change on polar bear populations."

Detail:

Include all necessary details to avoid ambiguity.

For instance, "Write a 500-word article on the economic impacts of climate change in developing countries."

2. Context Provision

Giving relevant background information and context within the prompt helps the AI understand the task and respond accurately.

This can include:

Background Information:

Provide the AI with necessary context. For example, "Given the recent IPCC report, summarize the key findings related to sea level rise."

Contextual Cues:

Use context to guide the AI. For instance, "As a historian, explain the significance of the Industrial Revolution."

3. Role Assignment (Role Prompting)

Assigning a specific role or persona to the AI model can guide the style and content of the response.

This technique is particularly useful for tailoring the output to specific needs:

Expert Roles: "You are a financial analyst. Explain the implications of the recent stock market crash."

Creative Roles: "You are a poet. Write a poem about the changing seasons."

4. Step-by-Step Instructions

Breaking down complex tasks into manageable steps can help the AI provide more structured and detailed responses:

Sequential Prompts: "First, list the ingredients needed for a chocolate cake. Then, describe the steps to bake it."

Process Explanation: "Explain how to set up a new email account, step by step."

5. Iterative Refinement

Continuously testing and refining the prompts based on the AI's responses is crucial for optimizing results:

Feedback Loop: Use the AI's output to refine your prompts. For example, if the response is too broad, narrow down the prompt.

Adjustments: Modify prompts based on the AI's performance. If the AI misunderstands a task, clarify the instructions in the next iteration.

6. Use of Examples

Providing example outputs or formats can clarify the expected response from the AI:

Sample Outputs: "Write a summary of this article. Example: 'This article discusses...'"

Format Guidance: "Generate a report in the following format: Introduction, Methods, Results, Conclusion."

7. Open-Ended vs. Closed-Ended Prompts

Choosing between prompts that encourage expansive answers or restrict responses to specific information can influence the quality of the output:

Open-Ended Prompts: "Describe the future of renewable energy." This allows for a broad and detailed response.

Closed-Ended Prompts: "List three benefits of renewable energy." This restricts the response to specific points.

8. Chain-of-Thought Prompts

Encouraging the AI to think through a problem step-by-step can improve the quality of complex responses:

Logical Sequencing: "To solve this math problem, first identify the variables, then set up the equations, and finally solve for the unknowns."

Reasoning: "Explain why the sky is blue, starting with the scattering of light."

9. Contextual Augmentation

Enhancing prompts with additional context or information can lead to more accurate and relevant responses:

Supplementary Information: "Considering the recent trends in the tech industry, predict the next big innovation."

Contextual Clues: "In the context of the 2020 pandemic, discuss the impact on global supply chains."

10. Meta-Prompts

Using prompts that guide the AI on how to handle the task itself can be very effective:

Instructional Prompts: "Explain the process of photosynthesis as if you are teaching a high school biology class."

Guidance Prompts: "When answering, consider the environmental, economic, and social factors."

11. Human-in-the-Loop

Involving human feedback in the prompt engineering process can enhance the quality of AI outputs:

Review and Feedback: Have a human review the AI's responses and provide feedback for improvement.

Iterative Collaboration: Work collaboratively with the AI, refining prompts based on human insights.

✍🏻 Prompt Frameworks

Here’s the MOST powerful thing you can do to get good results with prompts.

It’s using the right prompt frameworks.

Steal these frameworks 👇🏻

1/ RTF Prompt Framework

RTF stands for Role, Task, and Format.

And it’s a classic.

The role should be your current role, the task should be what you want it to accomplish, and the format should be the final output you want.

By default, I always resort to using this to build new prompts because of its ease of use.

Here’s a sample prompt.

2/ RACE Prompt Framework

RACE (Role, Action, Context, Expec is a slightly more advanced version of RTF. It gives you more flexibility to describe the context.

This is a great prompt framework for expanding further into specifics!

3/ RISEN Prompt Framework

RISEN is an advanced goal-based prompt framework.

By specifying steps to take, you will ensure that the AI comes up with ideas along your guardrails.

With specific end goals for work in mind, you will get a very focused scope from this prompt.

4/ RODES Prompt Framework

RODES is an advanced example-based prompt framework.

By specifying examples, it gives the LLM a frame of reference.

With such benchmarks, it is better able to offer similar strategies to what you specify.

Feel free to copy these prompt frameworks!

And… that’s it!

Stay tuned for more insights and advanced techniques in our upcoming newsletters.

Your feedback and questions are always welcome!

Best regards,

Justin 🍊

If you have any questions or feedback, reply to this email. (I'll personally reply! :)

Citations:

[1] https://www.datacamp.com/tutorial/a-beginners-guide-to-chatgpt-prompt-engineering

[2] https://www.coursera.org/articles/how-to-write-chatgpt-prompts

[3] https://help.openai.com/en/articles/6654000-best-practices-for-prompt-engineering-with-the-openai-api

[4] https://www.deeplearning.ai/short-courses/chatgpt-prompt-engineering-for-developers/

[5] https://www.zdnet.com/article/how-to-write-better-chatgpt-prompts-in-5-steps/

[6] https://www.digitalocean.com/resources/article/prompt-engineering-best-practices

[7] https://www.coursera.org/learn/prompt-engineering

[8] https://www.linkedin.com/pulse/how-craft-perfect-chatgpt-prompt-charlene-li-kowge

[9] https://cloud.google.com/blog/products/application-development/five-best-practices-for-prompt-engineering

[10] https://www.promptingguide.ai/models/chatgpt

[11] https://www.forbes.com/sites/jodiecook/2023/06/26/how-to-write-effective-prompts-for-chatgpt-7-essential-steps-for-best-results/

[12] https://platform.openai.com/docs/guides/prompt-engineering

[13] https://community.openai.com/t/chatgpt-prompt-engineering-for-developers/191028

[14] https://www.semrush.com/blog/chatgpt-prompts/

[15] https://www.techtarget.com/searchenterpriseai/tip/Prompt-engineering-tips-and-best-practices