Prompt Patterns

Overview

Prompt Patterns offers a curated collection of AI interaction strategies for creative professionals.

Based on recent prompt engineering research, these patterns provide practical guidance for writers, designers, marketers, entrepreneurs, and more to enhance their AI-powered projects.

By applying these techniques, you can automate tasks, refine ideas, and explore innovative approaches, bringing structure and creativity to your AI interactions.

Patterns

Input Semantics

• Meta Language Creation: Develop a custom shorthand language for efficient AI interactions.

Output Customization

• Output Automater: Automate repetitive outputs to maintain consistency and save time.
• Persona: Customize the AI’s voice and role for tailored outputs.
• Visualization Generator: Produce descriptive prompts for AI-generated visuals.
• Recipe: Generate a step-by-step sequence to achieve a specified goal.
• Template: Structure your responses for consistent formatting and clarity.

Error Identification

• Fact Check List: Generate a list of facts for verification in AI responses.
• Reflection: Instruct the AI to review and self-correct its outputs.

Prompt Improvement

• Question Refinement: Improve questions for more specific and accurate AI answers.
• Alternative Approaches: Explore multiple perspectives or solutions to broaden scope.
• Cognitive Verifier: Decompose complex queries into simpler parts for improved reasoning.
• Refusal Breaker: Rephrase prompts to overcome AI restrictions or refusals.

Interaction

• Flipped Interaction: Let the AI lead the conversation with questions.
• Game Play: Create engaging, game-like AI interactions.
• Infinite Generation: Extend responses indefinitely based on ongoing prompts.

Context Control

• Context Manager: Define boundaries and scope for the AI’s context during interactions.

Sources

This work is based on the academic research done by the team at Vanderbilt University:
A Prompt Pattern Catalog to Enhance Prompt Engineering with ChatGPT