The Art and Science of Prompt Engineering: Mastering Generative AI

Prompt engineering involves carefully crafting prompts to get generative AI models to produce the desired output. Mastering this skill requires understanding the model's capabilities and limitations.

Summary

• Assistant Professor Yeqing Kong shares three proven approaches to prompt engineering: the rhetorical approach, the C.R.E.A.T.E. framework, and the structured approach.
• The rhetorical approach involves stating your main claim and describing elements like audience, context, author ethos and credentials, emotional appeals, logic, arrangement, and style.
• The C.R.E.A.T.E. framework stands for Character, Request, Examples, Additions, Type of output, and Extras. You describe the AI's role, make a specific request, provide examples, add details, specify the output type, and share reference materials.
• The structured approach starts with the role, goal, and purpose; provides background context; defines the task; and supplies reference content.
• Prompt engineering takes iteration and refinement. Fix issues with confusing or unnecessary information to improve responses.
• Generative AI increases demand for liberal arts skills like writing, editing, and critical thinking that are key for prompt engineering.

READ MORE