Skip to content

Latest commit

 

History

History
37 lines (26 loc) · 5.41 KB

README.md

File metadata and controls

37 lines (26 loc) · 5.41 KB

VPAL logo

System Prompt Library

Concept and Need

The System Prompt Library is envisioned as a dynamic, user-centered digital repository designed for educators. Its primary purpose is to provide a comprehensive collection of customizable prompts for Generative Pre-trained Transformers (GPTs), tailored specifically for educational use.

System prompts play a critical role in the design of learning technologies enabled by generative artificial intelligence (GenAI). As part of a GenAI-powered tool for learning and instruction, they are responsible for guiding learner interaction, facilitating engagement and motivation, supporting pedagogical goals, providing feedback and assessment, ensuring content accuracy and relevance, encouraging reflective thinking and metacognition, facilitating adaptivity and personalization, and enhancing accessibility and inclusivity, among others. This site also outlines various strategies for designing effective system prompts, such as incorporating examples, character creation, narrative engagement, self-generated prompts, and sequential reasoning.

How to Use this System Prompt Library

This repository consists of several example system prompts, created and shared by Harvard-affiliated educators and by experts in the field of GenAI and education, separated into prompts specifically for educators and learners. For each prompt, we have created a brief summary of the prompt's purpose and utility, summarized the goals of the prompt in the form of specific learning objectives or instructional goals, provided a basic structure of the interaction each prompt affords, and curated relevant additional materials. Each section is intended to highlight the basic building blocks that go into system prompt design for educational purposes, and enable educators to experiment with, adapt, and create system prompts for their specific contexts.

To use a specific prompt, simply copy it from the "Copy This Prompt" code block and paste it into your custom GPT tool.

Each of the prompts found in this repository is licensed for reuse, however, licensing models have different requirements. Make sure to check the specific licensing requirements before copying, adapting, or sharing a specific system prompt.

Contribute to the System Prompt Library

We warmly welcome contributions from educators, developers, and educational technologists who are eager to enhance and expand this resource. If you're interested in contributing, here’s how you can get involved:

  1. Explore Existing Content: Start by browsing the current prompts in the library to understand the variety and scope of resources available.
  2. Identify Gaps or Opportunities: Think about what might be missing or how existing prompts could be improved. Do you have ideas for new prompts that could address specific educational challenges or goals?
  3. Fork the Repository: Visit our GitHub page and fork the repository to your own account. This will allow you to make changes without affecting the original project directly.
  4. Make Your Additions or Improvements: Add new prompts or update existing ones in your forked repository. Be sure to follow any guidelines provided for contributions to ensure consistency and quality.
  5. Submit a Pull Request: Once you’ve made your changes, submit a pull request. This is where you propose your changes to the original repository. Include a clear description of what you've done and why it adds value to the library.
  6. Review and Merge: The repository maintainers will review your pull request. If your contributions align with the library’s goals and quality standards, they will merge your changes into the main project.

If you are unfamiliar with GitHub you can submit a system prompt to be included in the library via our submission form:
Submit a system prompt

Disclaimer

It is important to remember that Large Language Models (LLMs) can sometimes respond in ways we didn't anticipate, which might be unsuitable or not what you expected. We advise you to tailor these prompts to match the specific needs and backgrounds of your students and your teaching context. Please make sure to check the LLMs' responses to ensure they're appropriate and accurate. Use your best judgment and be cautious when using these prompts in your teaching.

Who We Are

Nick Wilson, Director of Teaching Innovation
Dustin Tingley, Deputy Vice Provost for Advances in Teaching and Learning
Office of the Vice Provost for Advances in Learning
Harvard University