Created by Drake Rorabaugh - March 16, 2023

What makes a good prompt?

1. Specificity: Aim to make your prompts as specific as possible. Instead of asking "Tell me about climate change," you might ask "What are the primary causes of climate change according to current scientific consensus?"
2. Context: Providing context can help me give more appropriate responses. For example, if you're asking about the word 'run', clarify whether you mean it as a verb in a sports context, as a noun in a political context, or something else.
3. Directness: Be direct in your questions. If you need information on a specific subject, ask about it explicitly. For example, instead of saying "I'm interested in ancient civilizations," say "Can you give me information about the social structure of ancient Egypt?"
4. Open-Ended vs. Close-Ended: Decide whether you want an open-ended or close-ended answer. Open-ended questions (e.g., "What are some strategies for dealing with stress?") will elicit longer, more detailed responses, while close-ended questions (e.g., "Is Paris the capital of France?") result in brief, factual answers.
5. Intent: Make your intent clear. If you're looking for a debate, a list, a detailed explanation, or a quick fact-check, stating this explicitly can help guide the response.
6. Respect Constraints: It's important to understand that the language model has certain limitations. For instance, it doesn't have access to real-time data or updates beyond a certain cut-off date (as of the time of writing, the cut-off was in September 2021). Also, the model adheres to certain usage policies, such as not generating inappropriate content. Being aware of these constraints can help shape more effective prompts.
7. Iterate: If the first response isn't what you're looking for, try rephrasing or asking in a different way. Feedback is an important part of the process.

Not getting the responses you need? Try these methods.

• Craft Specific Questions: Ensure your questions are as specific and detailed as possible. Vague or broad questions can lead to generic or wide-ranging answers. Be clear about what you're asking for and provide as much context as you can.
• Resolve Misunderstandings: If the responses you receive don't seem to align with your question, there could be a misunderstanding. Try rephrasing your question or provide further explanation about what you're seeking.
• Experiment with Different Approaches: If you're not getting satisfactory results, don't hesitate to change your strategy. For instance, if you're asking for a list and the results are not as expected, try asking for one item at a time instead.
• Offer Constructive Feedback: Don't forget to share your observations if the responses aren't what you're looking for. Your feedback is valuable in refining and improving the system's responses.
• Remember, the goal of using ChatGPT is to foster effective communication and find the information you need. Be patient, persistent, and open to trying different strategies to achieve your desired results.

OpenAI’s Basics

• From OpenAI’s own guide: Prompt design Basics
  • Our models can do everything from generating original stories to performing complex text analysis. Because they can do so many things, you have to be explicit in describing what you want. Showing, not just telling, is often the secret to a good prompt.
  • There are three basic guidelines to creating prompts:
  • Show and tell. Make it clear what you want either through instructions, examples, or a combination of the two. If you want the model to rank a list of items in alphabetical order or to classify a paragraph by sentiment, show it that's what you want.
  • Provide quality data. If you're trying to build a classifier or get the model to follow a pattern, make sure that there are enough examples. Be sure to proofread your examples — the model is usually smart enough to see through basic spelling mistakes and give you a response, but it also might assume this is intentional and it can affect the response.
  • Check your settings. The temperature and top_p settings control how deterministic the model is in generating a response. If you're asking it for a response where there's only one right answer, then you'd want to set these lower. If you're looking for more diverse responses, then you might want to set them higher. The number one mistake people use with these settings is assuming that they're "cleverness" or "creativity" controls.

Prompts for ChatGPT

<aside> 🤖 Copy these prompts into a new chat: https://chat.openai.com/chat

</aside>