Prompt engineering is essentially asking artificial intelligence, or large language models to provide a desired answer for us. Prompts play a crucial role in guiding users and eliciting the desired responses. In this blog post, we will explore 20 different types of prompts that can be used for prompt engineering.

List of different types of prompts

Instructional Prompts

These prompts provide clear instructions to guide users on what action to take or how to proceed.

Here is an example of an instructional prompt in prompt engineering:

Prompt:

Write a poem about a cat, using the following words:
* furry
* playful
* curious
* cuddly
* mischievous

This prompt is instructional because it tells the language model what to write about (a cat) and what words to use (furry, playful, curious, cuddly, and mischievous). It also provides the language model with a clear goal (to write a poem).

Open-ended Prompts

These prompts encourage users to provide their own thoughts, ideas, or opinions without any specific constraints.

Here is an example of an open-ended prompt in prompt engineering:

Prompt:

Write a story about a character who discovers a new world.

This prompt is open-ended because it gives the language model a lot of freedom to be creative. The language model can choose any genre, setting, and characters it wants. It can also decide what kind of challenges the character faces and how they overcome them.

These prompts present users with a set of options to choose from, allowing them to select the most appropriate answer.

Fill-in-the-Blank Prompts

These prompts provide a partially completed sentence or phrase, prompting large language model to complete it with their own words.

Here is an example of a fill-in-the-blank prompt in prompt engineering:

The quick brown fox jumps over the lazy ___.

The LLM would be instructed to fill in the blank with a word or phrase that makes sense in the context of the sentence. In this case, the correct answer is “dog.”

Comparison Prompts

These prompts require users to compare two or more options and make a choice based on their preferences or criteria.

Prompt:

Compare and contrast the two following programming languages:

• Python
• Java

This prompt is also a comparison prompt because it asks the LLM to compare and contrast two different programming languages. The LLM is instructed to identify the similarities and differences between the two languages, and to explain how these similarities and differences affect the way that programs are written and executed.

Scenario-based Prompts

These prompts present users with a hypothetical situation or scenario and ask them to respond accordingly.

Here is an example of a scenario-based prompt in prompt engineering working with an LLM:

Prompt:

You are a customer service representative at a large e-commerce company. A customer has contacted you with a problem: they ordered a new phone, but it arrived damaged. The customer is angry and upset, and they want a full refund.

What would you say to the customer to try to resolve the issue?

This prompt is a scenario-based prompt because it presents the LLM with a hypothetical situation and asks it to respond in a way that would be appropriate for the situation. The LLM is instructed to take on the role of a customer service representative and to help the customer resolve their problem.

Reflection Prompts

These prompts encourage users to reflect on their thoughts, feelings, or experiences, promoting self-awareness.

Challenge Prompts

These prompts present language model with a problem or challenge, motivating them to think creatively and find a solution.

Here is an example of a challenge prompt in prompt engineering working with an LLM:

Prompt:

Generate a poem about a cat, using the following words:

• furry
• playful
• curious
• cuddly
• mischievous

This prompt is a challenge prompt because it requires the LLM to generate a poem that meets all of the following requirements:

• It must be about a cat.
• It must use the words furry, playful, curious, cuddly, and mischievous.
• It must be a poem.

This can be a challenging task for an LLM, but it is a good way to test its capabilities and to see what it is capable of creating.

14. Predictive Prompts: These prompts ask LLM to make predictions or guesses about future outcomes, encouraging critical thinking.

Prompt:

Write the next sentence in this story:

Once upon a time, there was a princess who lived in a tall tower. She was guarded by a dragon, and no one was allowed to see her.

This prompt is predictive because it asks the LLM to generate the next sentence in a story, based on the context of the previous sentence. The LLM must be able to understand the story so far and generate a sentence that makes sense and is consistent with the tone and style of the story.

Opinion Prompts

These prompts asks large language model to express their opinions on a particular subject, fostering dialogue.

Prompt:

What is your opinion on the following statement: “Artificial intelligence will eventually surpass human intelligence in all areas.”

This prompt is an opinion prompt because it asks the LLM to share its thoughts on a statement. The LLM is not instructed to generate a specific type of output but is instead given the freedom to express itself in whatever way it feels most comfortable.

Knowledge-based Prompts

These prompts test large language models’ knowledge on a specific topic or ask them to provide information.

Prompt:

Write a code snippet to calculate the area of a triangle, given the base and height of the triangle.

Problem-solving Prompts

These prompts present artificial intelligence (Large language model in this case) with a problem or puzzle to solve, stimulating their problem-solving skills.

Prompt:

How can we repair the bridge in the image above?

This prompt is problem-solving-based because it asks the LLM to come up with a solution to a problem. The LLM must be able to understand the problem and generate a solution that is both feasible and effective.

Conclusion

By leveraging the power of these different types of prompts, you can create dynamic and engaging experiences whilst working with large language models, this basic knowledge can be applied to any of the large language models. Whether that would be Bard, ChatGPT, or any other. Experiment with different prompt combinations and tailor them to your specific use case to maximize user interaction and satisfaction.

If you enjoyed this post, you can check our career guide to prompt engineering.