Prompt engineering: overview and guide
The rise of large language models (LLMs) has brought forth exciting possibilities for human-computer interaction. However, harnessing the full potential of these powerful AI models requires a crucial skill: prompt engineering. This burgeoning field focuses on crafting effective prompts that unlock the capabilities of LLMs, enabling them to understand intent, follow instructions, and generate desired outputs. As we increasingly interact with AI in various applications, prompt engineering plays a vital role in ensuring accurate, relevant, and safe interactions.
What is prompt engineering?
Prompt engineering is the art and science of designing and optimizing prompts to guide AI models, particularly LLMs, towards generating the desired responses. By carefully crafting prompts, you provide the model with context, instructions, and examples that help it understand your intent and respond in a meaningful way. Think of it as providing a roadmap for the AI, steering it towards the specific output you have in mind.
To dive deeper into the world of prompt design and explore its applications, check out the Introduction to Prompt Design on Google Cloud.
Ready to experiment with LLMs and prompt engineering firsthand? Try the Vertex AI free trial and experience the power of this technology.
What is a prompt for AI?
In the context of AI, a prompt is the input you provide to the model to elicit a specific response. This can take various forms, ranging from simple questions or keywords to complex instructions, code snippets, or even creative writing samples. The effectiveness of your prompt directly influences the quality and relevance of the AI's output.
What do you need for prompt engineering?
Several key elements contribute to effective prompt engineering. Mastering these allows you to communicate effectively with AI models and unlock their full potential.
Prompt format
The structure and style of your prompt play a significant role in guiding the AI's response. Different models may respond better to specific formats, such as:
The format of your prompt plays a significant role in how the AI interprets your request. Different models may respond better to specific formats, such as natural language questions, direct commands, or structured inputs with specific fields. Understanding the model's capabilities and preferred format is essential for crafting effective prompts.
Context and examples
Providing context and relevant examples within your prompt helps the AI understand the desired task and generate more accurate and relevant outputs. For instance, if you're looking for a creative story, including a few sentences describing the desired tone or theme can significantly improve the results.
Fine-tuning and adapting
Fine-tuning the AI model on specific tasks or domains using tailored prompts can enhance its performance. Additionally, adapting prompts based on user feedback or model outputs can further improve the model's responses over time.
Multi-turn conversations
Designing prompts for multi-turn conversations allows users to engage in continuous and context-aware interactions with the AI model, enhancing the overall user experience.
Types of prompts
There are various types of prompts used in AI, each serving a specific purpose:
Direct prompts (Zero-shot)
Zero-shot prompting involves providing the model with a direct instruction or question without any additional context or examples.
An example of this is idea generation, where the model is prompted to generate creative ideas or brainstorming solutions. Another example is summarization, or translation, where the model is asked to summarize or translate some piece of content.
One-, few- and multi-shot prompts
This method involves providing the model with one or more examples of the desired input-output pairs before presenting the actual prompt. This can help the model better understand the task and generate more accurate responses.
Chain of Thought Prompts
CoT prompting encourages the model to break down complex reasoning into a series of intermediate steps, leading to a more comprehensive and well-structured final output.
Zero-shot CoT Prompts
Combines chain of thought prompting with zero-shot prompting by asking the model to perform reasoning steps, which may often produce better output.
Use cases and examples of prompt engineering
Here are some specific examples and use cases showing how prompt engineering helps produce customized and relevant output.
Language and Text Generation
Scenario | Example Prompt | |
Creative Writing | Craft prompts that specify genre, tone, style, and plot points to guide the AI in generating engaging narratives. | "Write a short story about a young woman who discovers a magical portal in her attic." |
Summarization | Provide the AI with text and instruct it to generate concise summaries that capture key information. | “Summarize the main points of the following news article on climate change." |
Translation | Specify the source and target languages to enable the AI to accurately translate text while preserving meaning and context. | "Translate the following text from English to Spanish: 'The quick brown fox jumps over the lazy dog.'" |
Dialogue | Design prompts that simulate conversations, allowing the AI to generate responses that mimic human interaction and maintain context. | "You are a friendly chatbot helping users troubleshoot their computer problems. Respond to the user's query: 'My computer won't turn on.'" |
Scenario
Example Prompt
Creative Writing
Craft prompts that specify genre, tone, style, and plot points to guide the AI in generating engaging narratives.
"Write a short story about a young woman who discovers a magical portal in her attic."
Summarization
Provide the AI with text and instruct it to generate concise summaries that capture key information.
“Summarize the main points of the following news article on climate change."
Translation
Specify the source and target languages to enable the AI to accurately translate text while preserving meaning and context.
"Translate the following text from English to Spanish: 'The quick brown fox jumps over the lazy dog.'"
Dialogue
Design prompts that simulate conversations, allowing the AI to generate responses that mimic human interaction and maintain context.
"You are a friendly chatbot helping users troubleshoot their computer problems. Respond to the user's query: 'My computer won't turn on.'"
Question Answering
Scenario | Instructions | Example Prompt |
Open-Ended Questions | Formulate prompts that encourage the AI to provide comprehensive and informative answers based on its knowledge base. | "Explain the concept of quantum computing and its potential impact on the future of technology." |
Specific Questions | Design prompts that target specific information, enabling the AI to retrieve precise answers from the provided context or its internal knowledge base. | "What is the capital of France?" or "According to the provided text, what are the main causes of deforestation?" |
Multiple Choice Questions | Present prompts with options, prompting the AI to analyze and select the most appropriate answer based on its understanding of the context. | "Who wrote the Harry Potter series? A) J.R.R. Tolkien, B) J.K. Rowling, C) Stephen King" |
Hypothetical Questions | Craft prompts that explore hypothetical situations, allowing the AI to reason, speculate, and provide potential outcomes or solutions. | "What would happen if humans could travel at the speed of light?" |
Opinion-Based Questions | Design prompts that elicit the AI's perspective or opinion on a specific topic, encouraging it to provide reasoning and justification for its stance. | "Do you believe that artificial intelligence will eventually surpass human intelligence? Why or why not?" |
Scenario
Instructions
Example Prompt
Open-Ended Questions
Formulate prompts that encourage the AI to provide comprehensive and informative answers based on its knowledge base.
"Explain the concept of quantum computing and its potential impact on the future of technology."
Specific Questions
Design prompts that target specific information, enabling the AI to retrieve precise answers from the provided context or its internal knowledge base.
"What is the capital of France?" or "According to the provided text, what are the main causes of deforestation?"
Multiple Choice Questions
Present prompts with options, prompting the AI to analyze and select the most appropriate answer based on its understanding of the context.
"Who wrote the Harry Potter series? A) J.R.R. Tolkien, B) J.K. Rowling, C) Stephen King"
Hypothetical Questions
Craft prompts that explore hypothetical situations, allowing the AI to reason, speculate, and provide potential outcomes or solutions.
"What would happen if humans could travel at the speed of light?"
Opinion-Based Questions
Design prompts that elicit the AI's perspective or opinion on a specific topic, encouraging it to provide reasoning and justification for its stance.
"Do you believe that artificial intelligence will eventually surpass human intelligence? Why or why not?"
Code Generation
Scenario | Instructions | Example Prompt |
Code Completion | Provide the AI with a partial code snippet and prompt it to suggest or complete the remaining code based on the context and programming language. | "Write a Python function to calculate the factorial of a given number." |
Code Translation | Specify the source and target programming languages to enable the AI to translate code while preserving functionality and syntax. | "Translate the following Python code to JavaScript: def greet(name): print('Hello,', name)" |
Code Optimization | Prompt the AI to analyze existing code and suggest improvements for efficiency, readability, or performance. | "Optimize the following Python code to reduce its execution time." |
Code Debugging | Provide the AI with code containing errors and prompt it to identify and suggest potential solutions for the identified issues. | "Debug the following Java code and explain why it is throwing a NullPointerException."
|
Scenario
Instructions
Example Prompt
Code Completion
Provide the AI with a partial code snippet and prompt it to suggest or complete the remaining code based on the context and programming language.
"Write a Python function to calculate the factorial of a given number."
Code Translation
Specify the source and target programming languages to enable the AI to translate code while preserving functionality and syntax.
"Translate the following Python code to JavaScript: def greet(name): print('Hello,', name)"
Code Optimization
Prompt the AI to analyze existing code and suggest improvements for efficiency, readability, or performance.
"Optimize the following Python code to reduce its execution time."
Code Debugging
Provide the AI with code containing errors and prompt it to identify and suggest potential solutions for the identified issues.
"Debug the following Java code and explain why it is throwing a NullPointerException."
Image Generation
Scenario | Instructions | Example Prompt |
Photorealistic Images | Craft prompts that describe the desired image in detail, including objects, scenery, lighting, and style, to generate realistic and high-quality images. | "A photorealistic image of a sunset over the ocean with palm trees silhouetted against the sky." |
Artistic Images | Design prompts that specify art styles, techniques, and subject matter to guide the AI in creating images that mimic specific artistic movements or evoke certain emotions. | "An impressionist painting of a bustling city street with people walking under umbrellas in the rain." |
Abstract Images | Formulate prompts that encourage the AI to generate images that are open to interpretation, utilizing shapes, colors, and textures to evoke feelings or concepts. | "An abstract image representing the concept of hope, using bright colors and flowing shapes." |
Image Editing | Provide the AI with an existing image and specify desired modifications, enabling it to edit and enhance the image according to the given instructions. | "Change the background of this photo to a starry night sky and add a full moon." or "Remove the person from this image and replace them with a cat." |
Scenario
Instructions
Example Prompt
Photorealistic Images
Craft prompts that describe the desired image in detail, including objects, scenery, lighting, and style, to generate realistic and high-quality images.
"A photorealistic image of a sunset over the ocean with palm trees silhouetted against the sky."
Artistic Images
Design prompts that specify art styles, techniques, and subject matter to guide the AI in creating images that mimic specific artistic movements or evoke certain emotions.
"An impressionist painting of a bustling city street with people walking under umbrellas in the rain."
Abstract Images
Formulate prompts that encourage the AI to generate images that are open to interpretation, utilizing shapes, colors, and textures to evoke feelings or concepts.
"An abstract image representing the concept of hope, using bright colors and flowing shapes."
Image Editing
Provide the AI with an existing image and specify desired modifications, enabling it to edit and enhance the image according to the given instructions.
"Change the background of this photo to a starry night sky and add a full moon." or "Remove the person from this image and replace them with a cat."
Strategies for writing better prompts
Developing effective prompts requires a strategic approach. Consider these strategies to enhance your prompt engineering skills:
1. Set Clear Goals and Objectives:
Tactic | Prompt Example |
Use action verbs to specify the desired action | "Write a bulleted list that summarizes the key findings of the attached research paper" |
Define the desired length and format of the output | "Compose a 500-word essay discussing the impact of climate change on coastal communities." |
Specify the target audience | "Write a product description for a new line of organic skincare products, targeting young adults concerned with sustainability." |
Tactic
Prompt Example
Use action verbs to specify the desired action
"Write a bulleted list that summarizes the key findings of the attached research paper"
Define the desired length and format of the output
"Compose a 500-word essay discussing the impact of climate change on coastal communities."
Specify the target audience
"Write a product description for a new line of organic skincare products, targeting young adults concerned with sustainability."
2. Provide Context and Background Information:
Tactic | Prompt Example |
Include relevant facts and data | "Given that global temperatures have risen by 1 degree Celsius since the pre-industrial era, discuss the potential consequences for sea level rise." |
Reference specific sources or documents | "Based on the attached financial report, analyze the company's profitability over the past five years." |
Define key terms and concepts | "Explain the concept of quantum computing in simple terms, suitable for a non-technical audience." |
Tactic
Prompt Example
Include relevant facts and data
"Given that global temperatures have risen by 1 degree Celsius since the pre-industrial era, discuss the potential consequences for sea level rise."
Reference specific sources or documents
"Based on the attached financial report, analyze the company's profitability over the past five years."
Define key terms and concepts
"Explain the concept of quantum computing in simple terms, suitable for a non-technical audience."
3. Use Few-Shot Prompting:
Tactic | Prompt Example |
Provide a few examples of desired input-output pairs | Input: "Cat" Output: "A small furry mammal with whiskers." Input: "Dog" Output: "A domesticated canine known for its loyalty." Prompt: "Elephant" |
Demonstrate the desired style or tone | Example 1 (humorous): "The politician's speech was so dull, it could cure insomnia." Example 2 (formal): "The dignitary delivered an address that was both informative and engaging." Prompt: "Write a sentence describing the comedian's stand-up routine." |
Show the desired level of detail | Example 1 (brief): "The movie was about a young boy who befriends an alien." Example 2 (detailed): "The science fiction film follows the story of Elliot, a lonely boy who discovers and forms a unique bond with an extraterrestrial stranded on Earth." Prompt: "Summarize the plot of the novel you just finished reading." |
Tactic
Prompt Example
Provide a few examples of desired input-output pairs
Input: "Cat" Output: "A small furry mammal with whiskers." Input: "Dog" Output: "A domesticated canine known for its loyalty." Prompt: "Elephant"
Demonstrate the desired style or tone
Example 1 (humorous): "The politician's speech was so dull, it could cure insomnia." Example 2 (formal): "The dignitary delivered an address that was both informative and engaging." Prompt: "Write a sentence describing the comedian's stand-up routine."
Show the desired level of detail
Example 1 (brief): "The movie was about a young boy who befriends an alien." Example 2 (detailed): "The science fiction film follows the story of Elliot, a lonely boy who discovers and forms a unique bond with an extraterrestrial stranded on Earth." Prompt: "Summarize the plot of the novel you just finished reading."
4. Be Specific:
Tactic | Prompt Example |
Use precise language and avoid ambiguity | Instead of: "Write something about climate change," use: "Write a persuasive essay arguing for the implementation of stricter carbon emission regulations." |
Quantify your requests whenever possible | Instead of: "Write a long poem," use: "Write a sonnet with 14 lines that explores themes of love and loss." |
Break down complex tasks into smaller steps | Instead of: "Create a marketing plan," use: "1. Identify the target audience. 2. Develop key marketing messages. 3. Choose appropriate marketing channels." |
Tactic
Prompt Example
Use precise language and avoid ambiguity
Instead of: "Write something about climate change," use: "Write a persuasive essay arguing for the implementation of stricter carbon emission regulations."
Quantify your requests whenever possible
Instead of: "Write a long poem," use: "Write a sonnet with 14 lines that explores themes of love and loss."
Break down complex tasks into smaller steps
Instead of: "Create a marketing plan," use: "1. Identify the target audience. 2. Develop key marketing messages. 3. Choose appropriate marketing channels."
5. Iterate and Experiment:
Tactic | Action |
Try different phrasings and keywords | Rephrase your prompt using synonyms or alternative sentence structures. |
Adjust the level of detail and specificity | Add or remove information to fine-tune the output. |
Test different prompt lengths | Experiment with both shorter and longer prompts to find the optimal balance. |
Tactic
Action
Try different phrasings and keywords
Rephrase your prompt using synonyms or alternative sentence structures.
Adjust the level of detail and specificity
Add or remove information to fine-tune the output.
Test different prompt lengths
Experiment with both shorter and longer prompts to find the optimal balance.
6. Leverage Chain of Thought Prompting:
Tactic | Prompt Example |
Encourage step-by-step reasoning | "Solve this problem step-by-step: John has 5 apples, he eats 2. How many apples does he have left? Step 1: John starts with 5 apples. Step 2: He eats 2 apples, so we need to subtract 2 from 5. Step 3: 5 - 2 = 3. Answer: John has 3 apples left." |
Ask the model to explain its reasoning process | "Explain your thought process in determining the sentiment of this movie review: 'The acting was superb, but the plot was predictable.'" |
Guide the model through a logical sequence of thought | "To classify this email as spam or not spam, consider the following: 1. Is the sender known? 2. Does the subject line contain suspicious keywords? 3. Is the email offering something too good to be true?" |
Tactic
Prompt Example
Encourage step-by-step reasoning
"Solve this problem step-by-step: John has 5 apples, he eats 2. How many apples does he have left? Step 1: John starts with 5 apples. Step 2: He eats 2 apples, so we need to subtract 2 from 5. Step 3: 5 - 2 = 3. Answer: John has 3 apples left."
Ask the model to explain its reasoning process
"Explain your thought process in determining the sentiment of this movie review: 'The acting was superb, but the plot was predictable.'"
Guide the model through a logical sequence of thought
"To classify this email as spam or not spam, consider the following: 1. Is the sender known? 2. Does the subject line contain suspicious keywords? 3. Is the email offering something too good to be true?"
For further guidance on prompt engineering best practices, explore the Five Best Practices for Prompt Engineering on Google Cloud.
Benefits of prompt engineering
Effective prompt engineering offers numerous benefits, enhancing the capabilities and usability of AI models:
Improved model performance
Well-crafted prompts lead to more accurate, relevant, and informative outputs from AI models, as they provide clear instructions and context.
Reduced bias and harmful responses
By carefully controlling the input and guiding the AI's focus, prompt engineering helps mitigate bias and minimize the risk of generating inappropriate or offensive content.
Increased control and predictability
Prompt engineering empowers you to influence the AI's behavior and ensure consistent and predictable responses aligned with your desired outcomes.
Enhanced user experience
Clear and concise prompts make it easier for users to interact effectively with AI models, leading to more intuitive and satisfying experiences.
Start your AI journey with Google Cloud
Related Google Cloud products and services
Vertex AI PlatformA single platform for data scientists and engineers to create, train, test, monitor, tune, and deploy ML and AI models.- Generative AI on Vertex AIRapidly prototype and test generative AI models. Test sample prompts, design your own prompts, and customize foundation models and LLMs.
AI APIsEasily integrate AI into your applications with Google Cloud's AI and machine learning APIs.Solution
Model Garden on Vertex AIJumpstart your ML project with a single place to discover, customize, and deploy a wide variety of models from Google and Google partners.
Additional learning resources to get started
New to Google Cloud or generative AI? New customers get $300 in free credits to run, test, and deploy workloads.
Take the next step
Start building on Google Cloud with $300 in free credits and 20+ always free products.
Need help getting started?
Contact salesWork with a trusted partner
Find a partnerContinue browsing
See all products
