Prompt engineering is the process of designing and refining the inputs given to language models, like those in AI, to achieve desired outputs more effectively. It involves creatively crafting prompts that guide the model in generating responses that are accurate, relevant, and aligned with the user's intentions.
For students and researchers in higher education, mastering prompt engineering is essential for leveraging the full potential of AI in research, development, and practical applications.
| Number | Principle | Description | Example |
| 1 | Give Direction | Clearly describe the desired outcome, style, or reference a relevant academic persona to guide the model's output. | "Write an outline of recent advancements in nanotechnology for drug delivery systems in the style of a review article for 'Nature Nanotechnology'." "Draft an essay outline discussing the ethical implications of artificial intelligence in surveillance, ensuring to reflect upon perspectives from both privacy advocates and security experts. The essay should emulate the analytical depth found in the 'Harvard Journal of Law & Technology'." |
| 2 | Specify Format | Detail the rules to follow and the structure the response should take, such as APA format for citations or a structured abstract. | "Create a structured abstract for a research paper on machine learning in genetic sequencing, including sections for Objective, Methods, Results, and Conclusion, following APA guidelines." "Construct a conference paper outline on the development of smart cities, including sections for Introduction, Background, Methodology, Results, Discussion, and Conclusion, adhering to IEEE formatting guidelines." |
| 3 | Provide Examples | Incorporate a diverse set of examples or test cases to illustrate how the task should be correctly executed, enhancing the model's understanding. | "Given examples of successful AI applications in sustainable agriculture, summarize how each application impacts resource efficiency and crop yield." "Review instances of machine learning algorithms being applied to predict stock market trends. Outline the methodology, data sources, and accuracy of predictions for each instance, noting any improvements over traditional models." |
| 4 | Evaluate Quality | Critically identify errors, rate responses, and conduct tests to understand what factors contribute to the model's performance and accuracy. | "After generating a literature review on the use of AI in diagnosing rare diseases, evaluate the completeness, relevance, and accuracy of cited studies." "Conduct a comparative analysis of the energy efficiency of three different types of solar panels. Include in your evaluation the cost-benefit ratio, lifespan, and performance under various weather conditions, supported by recent research findings. Provide the results in a table.” |
| 5 | Chain of Thought | Break down complex tasks into multiple, smaller steps, chaining them together to achieve comprehensive and nuanced goals. |
"First, identify key areas of interest in renewable energy research. Next, generate a series of questions that explore these areas. Finally, compile a list of potential methodologies to investigate these questions." “Perform the following steps in a consecutive order: Step 1, Step 2, Step 3, Step 4, and Step 5. Step 1 – Generate an answer for the 3 most popular sub-topics related to "autonomous vehicles and natural disasters". Step 2 – Generate 3 of the most popular sub-topics related to each answer. Step 3 – Take those answers and list their top 3 most popular keywords without a description. Step 4 – For the answers given of their most popular keywords, provide 3 long-tail keywords. Step 5 – for each long-tail keyword offered in the response, a list without descriptions 3 of their top semantically related keywords and entities.” |