GenaiHow-ToBeginner · 4 min read

How to Use Examples in Prompts for Better AI Responses

To use examples in prompts, include clear input-output pairs that show the AI what you want it to do. This helps the model understand the task better and produce accurate results by following the pattern you provide.

📐

Syntax

Using examples in prompts means giving the AI a few sample inputs and their correct outputs before asking it to generate a new output. This is often called few-shot prompting.

The basic syntax looks like this:

Example 1 Input: The first sample input.
Example 1 Output: The correct output for the first input.
Example 2 Input: The second sample input.
Example 2 Output: The correct output for the second input.
New Input: The input you want the AI to answer.
New Output: The AI's response based on the examples.

This pattern teaches the AI how to respond by showing it examples first.

plaintext

prompt = '''
Input: Translate 'hello' to French.
Output: bonjour

Input: Translate 'thank you' to French.
Output: merci

Input: Translate 'goodbye' to French.
Output: '''

💻

Example

This example shows how to use examples in a prompt to translate English words to French. The AI sees two examples and then translates a new word.

python

def generate_prompt(word):
    prompt = f"""
Input: Translate 'hello' to French.
Output: bonjour

Input: Translate 'thank you' to French.
Output: merci

Input: Translate '{word}' to French.
Output:"""
    return prompt

# Example usage
word_to_translate = 'goodbye'
prompt_text = generate_prompt(word_to_translate)
print(prompt_text)

Output

Input: Translate 'hello' to French. Output: bonjour Input: Translate 'thank you' to French. Output: merci Input: Translate 'goodbye' to French. Output:

⚠️

Common Pitfalls

Common mistakes when using examples in prompts include:

Using unclear or inconsistent examples that confuse the AI.
Providing too few or too many examples, which can underfit or overwhelm the model.
Not matching the format exactly between examples and the new input.
Forgetting to leave the output blank for the new input so the AI knows to generate it.

Here is a wrong and right way to write examples:

plaintext

# Wrong way: Missing output label for new input
wrong_prompt = '''
Input: Add 2 + 2.
Output: 4

Input: Add 3 + 5.
Output: 8

Input: Add 7 + 1.
'''

# Right way: Output label included for AI to complete
right_prompt = '''
Input: Add 2 + 2.
Output: 4

Input: Add 3 + 5.
Output: 8

Input: Add 7 + 1.
Output: '''

📊

Quick Reference

Tips for using examples in prompts:

Keep examples clear and consistent in format.
Use 2-5 examples to guide the AI well.
Always leave the output blank for the new input.
Match the style and wording exactly between examples and new input.
Test and adjust examples if the AI output is off.

✅

Key Takeaways

Use clear input-output pairs as examples to guide AI responses.

Keep example formats consistent and leave new output blank for AI to complete.

Provide a few examples (2-5) to balance guidance without overload.

Avoid mixing formats or unclear examples to prevent confusing the AI.

Test prompts and refine examples for best results.