A Sample example of:

Prompt Engineering best practices.

 

Here is a sample example illustrating some prompt engineering best practices:

Scenario: You want to generate a short product description for a new type of smartwatch aimed at fitness enthusiasts.

Ineffective Prompt: "Write a product description for a smartwatch."

This prompt is too vague and doesn't provide enough context for the model to generate a useful and targeted description.

Better Prompt: "Here are some key features of the new FitTech PowerWatch for fitness enthusiasts:

• Continuous heart rate monitoring
• GPS and workout tracking for running, cycling, hiking etc.
• Waterproof up to 50m for swimming
• 5-day battery life
• Lightweight at just 40g

Using those details, write a compelling and concise product description for the FitTech PowerWatch, highlighting its key fitness and activity tracking capabilities. Keep the description around 4-5 sentences long."

This improved prompt:

1. Sets the context by providing relevant product details and the target audience.
2. Gives an example of the desired output format/length.
3. Specifies the tone (compelling, concise) and focus (fitness/activity).
4. Provides clear instructions on what to include in the description.

This allows the model to generate a more tailored and useful product description, such as:

"The new FitTech PowerWatch is the ultimate fitness companion. Stay motivated with continuous heart rate tracking, GPS mapping for all your workouts, and advanced activity metrics - whether running, cycling or hiking. This featherweight smartwatch is waterproof to 50m for swim tracking yet lasts up to 5 days on a single charge. Power through your fitness goals effortlessly."

By including the right prompt components like context, examples, instructions, and output specifications, you can greatly improve the quality and relevance of model-generated text. Iterating prompts with techniques like this is considered a prompt engineering best practice.