Jules White and colleagues at Vanderbilt University have created a catalog of prompt patterns to improve interactions with large language models like ChatGPT. These patterns, similar to design patterns in software engineering, offer structured techniques for crafting prompts to obtain predictable and beneficial responses from AI systems. The patterns are organized into six categories: Input Semantics, Output Customization, Error Identification, Prompt Improvement, Interaction, and Context Control. Each pattern includes a definition, example, and use case illustrating its practical application across fields like marketing, education, and software development. By using these patterns, users can enhance AI's capabilities for various tasks, including automation, content generation, problem-solving, and decision-making. Ultimately, the catalog aims to make AI interactions more precise, reliable, and useful through the strategic design of prompts.

Ver todos los episodios