Custom GPT prompt instruction checklist
1. Understanding the Objective
- Describe the demographic characteristics.
- Understand the knowledge level of the audience.
- Consider their interests and preferences.
- Identify any specific needs or challenges.
- Tailor the language and tone accordingly.
- Specify the nature of the content needed.
- Decide on the format (article, dialogue, report).
- Outline any constraints or requirements.
- Consider the intended use of the content.
- Ensure alignment with overall project goals.
2. Structuring the Prompt
- Avoid jargon and complex terminology.
- Use simple words and short sentences.
- Be direct and to the point.
- Eliminate unnecessary details.
- Ensure clarity to enhance understanding.
- Clearly indicate the desired output format.
- Use labels such as 'List' or 'Paragraph'.
- Specify if code snippets are needed.
- Detail any specific formatting requirements.
- Provide examples of the format if necessary.
- Provide at least one example for clarity.
- Use relevant and relatable scenarios.
- Ensure examples match the prompt context.
- Include both good and bad examples if applicable.
- Keep examples concise and focused.
3. Contextual Information
- Identify the main topic or subject area.
- Summarize key concepts that influence the prompt.
- Detail any relevant industry standards or practices.
- Explain the significance of the context to the prompt.
- Specify word count limits for responses.
- Outline any prohibited topics or areas.
- Indicate preferred formatting styles or structures.
- Define the scope of acceptable responses.
- List any foundational concepts the model should understand.
- Specify relevant terminology or jargon.
- Indicate any previous interactions or discussions.
- Clarify assumptions about the audience's expertise.
4. Refining the Instructions
- Write instructions using clear, action-oriented verbs.
- Avoid passive constructions that may cause confusion.
- Ensure commands are concise and straightforward.
- Example: 'Generate a report' instead of 'A report should be generated.'
- Define all terms and concepts explicitly.
- Use specific language to eliminate multiple interpretations.
- Provide context where necessary to clarify references.
- Example: Instead of 'it', specify the subject.
- Conduct reviews with diverse team members.
- Gather feedback to identify potential misinterpretations.
- Simulate various user scenarios to evaluate responses.
- Iterate on the prompt based on testing results.
5. Iterative Testing and Feedback
6. Documentation and Version Control
- Use a standardized format for documentation.
- Include the prompt's purpose and intended use.
- Store the document in a shared location accessible to the team.
- Regularly review and update the documentation as needed.
- Create a version control system, such as Git.
- Log each change with a date and brief description.
- Tag releases for significant updates.
- Ensure all team members understand how to use the version control system.
- Encourage team members to provide feedback.
- Document successes and challenges in a dedicated section.
- Regularly review notes during team meetings.
- Prioritize enhancements based on team input and user feedback.
7. Compliance and Ethical Considerations
- Review relevant industry ethical guidelines.
- Consult legal frameworks applicable to your domain.
- Conduct a risk assessment for potential ethical breaches.
- Document compliance measures and decisions.
- Identify potentially sensitive topics related to the prompt.
- Evaluate the necessity of including controversial subjects.
- Provide context and disclaimers when necessary.
- Monitor audience reactions and feedback for adjustments.
- Use gender-neutral language wherever possible.
- Incorporate diverse perspectives in examples.
- Avoid stereotypes and generalizations in wording.
- Seek feedback from diverse groups to improve inclusivity.
8. Deployment and Monitoring
- Integrate the prompt into the application codebase.
- Ensure compatibility with existing systems and frameworks.
- Conduct initial tests to validate functionality.
- Deploy the application to a staging environment first.
- Confirm deployment success before moving to production.
- Set up analytics to track user engagement and prompt usage.
- Collect qualitative feedback from users regularly.
- Evaluate response times and accuracy of the prompt.
- Review logs for unexpected errors or issues.
- Adjust metrics to refine monitoring as needed.
- Analyze feedback to identify common issues or suggestions.
- Prioritize changes based on user impact and feasibility.
- Implement updates in a test environment before deployment.
- Communicate changes to users to manage expectations.
- Schedule regular reviews for ongoing improvements.
Use in Manifestly