1 min read Learning Center

4. Unlock the Power of Prompt Engineering with the "BROKE"

Today, let’s dive into something fascinating: the BROKE principle for prompt engineering. Don’t worry, the name doesn’t mean it’s flawed. It’s actually a handy mnemonic to guide us in crafting great prompts for AI models like ChatGPT.

So, what does BROKE stand for? Let me break it down for you:

B: Be Specific

When you ask something vague, you often get vague answers. A specific prompt leads to specific, high-quality results.

  • Example: Instead of asking, “Tell me about history,” try “What are the main causes of World War II?”

R: Role Assignment

AI performs better when given a role to play. This helps the model tailor its tone, style, and information.

  • Example: Ask, “You are a history professor. Can you explain the causes of World War II?”

O: Output Format

Let the AI know how you want the information structured. This makes it easier to use the response.

  • Example: Request, “Provide a bullet-point summary of World War II causes.”

K: Knowledge Context

Provide background or extra details to help the AI understand your request better.

  • Example: Mention, “I am preparing for a high school history exam and need clear, concise information about World War II.”

E: Examples

Examples act as a template for the AI, helping it mimic the style or structure you want.

  • Example: Add, “Here’s an example of a summary: [Insert your example]. Could you write something similar?”

By following BROKE, you can turn a plain prompt into a powerful one, boosting the quality and relevance of your AI responses.

yicong qin