It’s been a few months since ChatGPT released its GPT feature. This allows users with Plus accounts (paying users) to create customized GPTs. You can learn more about it from OpenAI here. There seems to be a lot of hype around these but that seemed to fade for most folks.
I’ve played around with these the last few months and haven’t found them to be particularly exciting. For instance, I created one that loaded all of my notes, guidelines, support materials from my Popular Culture course to create a kind of “Pop Culture GPTutor”. After loading and playing with it for a few hours, I found that it didn’t offer much use. It failed more often than helped in terms of its responses. It couldn’t answer in ways that I felt would be helpful even after continuing to adjust the tone. I’ve seen a few others GPTs and been not excited by them, to say the least.
I tried to create a Research Review GPT for my series on exploring research using generative AI, but that too was actually insufficient and produced lesser results than just starting a new thread. So if felt like something that I stopped playing with them for a few weeks.
But recently, I decided to give it one more try. In my feeds, an article popped up last week that’s making the rounds: Principled Instructions Are All You Need for Questioning LLaMA-1/2, GPT-3.5/4 by Sondos Mahmoud Bsharat, Aidar Myrzakhan, and Zhiqiang Shen. The authors have put together a rather substantial and clear list of 26 guidance and instructions that are helpful in eliciting better responses from large language models. They include the following:
1. No need to be polite with LLM so there is no need to add phrases like “please”, “if you don’t mind”, “thank you”, “I would like to”, etc., and get straight to the point.
2. Integrate the intended audience in the prompt, e.g., the audience is an expert in the field.
3. Break down complex tasks into a sequence of simpler prompts in an interactive conversation.
4. Employ affirmative directives such as ‘do,’ while steering clear of negative language like ‘don’t’.
5. When you need clarity or a deeper understanding of a topic, idea, or any piece of information, utilize the following prompts:
Explain [insert specific topic] in simple terms.
Explain to me like I’m 11 years old.
Explain to me as if I’m a beginner in [field].
Write the [essay/text/paragraph] using simple English like you’re explaining something to a 5-year-old.
6. Add “I’m going to tip $xxx for a better solution!”
7. Implement example-driven prompting (Use few-shot prompting).
8. When formatting your prompt, start with ‘###Instruction###’, followed by either ‘###Example###’ or ‘###Question###’ if relevant. Subsequently, present your content. Use one or more line breaks to separate instructions, examples, questions, context, and input data.
9. Incorporate the following phrases: “Your task is” and “You MUST”.
10. Incorporate the following phrases: “You will be penalized”.
11. Use the phrase ”Answer a question given in a natural, human-like manner” in your prompts.
12. Use leading words like writing “think step by step”.
13. Add to your prompt the following phrase “Ensure that your answer is unbiased and does not rely on stereotypes”.
14. Allow the model to elicit precise details and requirements from you by asking you questions until he has enough information to provide the needed output (for example, “From now on, I would like you to ask me questions to...”).
15. To inquire about a specific topic or idea or any information and you want to test your understanding, you can use the following phrase: “Teach me the [Any theorem/topic/rule name] and include a test at the end, but don’t give me the answers and then tell me if I got the answer right when I respond”.
16. Assign a role to the large language models.
17. Use Delimiters.
18. Repeat a specific word or phrase multiple times within a prompt.
19. Combine Chain-of-thought (CoT) with few-Shot prompts.
20. Use output primers, which involve concluding your prompt with the beginning of the desired output. Utilize output primers by ending your prompt with the start of the anticipated response.
21. To write an essay /text /paragraph /article or any type of text that should be detailed: “Write a detailed [essay/text /paragraph] for me on [topic] in detail by adding all the information necessary”.
22. To correct/change specific text without changing its style: “Try to revise every paragraph sent by users. You should only improve the user’s grammar and vocabulary and make sure it sounds natural. You should not change the writing style, such as making a formal paragraph casual”.
23. When you have a complex coding prompt that may be in different files: “From now and on whenever you generate code that spans more than one file, generate a [programming language ] script that can be run to automatically create the specified files or make changes to existing files to insert the generated code. [your question]”.
24. When you want to initiate or continue a text using specific words, phrases, or sentences, utilize the following prompt:
I’m providing you with the beginning [song lyrics/story/paragraph/essay...]: [Insert lyrics/words/sentence]’.
Finish it based on the words provided. Keep the flow consistent.
25. Clearly state the requirements that the model must follow in order to produce content, in the form of the keywords, regulations, hint, or instructions
26. To write any text, such as an essay or paragraph, that is intended to be similar to a provided sample, include the following instructions:
Please use the same language based on the provided paragraph[/title/text /essay/answer
Many of these I am somewhat familiar with and others are new; still some I need to look up to better understand myself. But this gave me a thought. How could I regularly used these to get better responses generative AI.
I’m getting better at asking good questions that limit the amount of interactions with generative AI but I still stumble a lot or have to ask the same thing a few times to get the right response. So what if I had a go-to that could help me improve upon the prompt?
Using that article as part of its context, I create this Prompt Enhance GPT. Its purpose is to enhance prompts based upon the principles above. Once I have a prompt, I take it there first and ask it
Review, revise, and enhance this prompt so that a large-language model provides its most creative and deepest-thinking results:
[Initial Prompt]
It will review the prompt, and then provide some commentary. Then, it rewrites the prompt accordingly. Sometimes, that prompt will have placeholder brackets for certain pieces of information that I need to supply.
Now, here’s the trick. I take that prompt and do it one more time.
Review, revise, and enhance this prompt so that a large-language model provides its most creative and deepest-thinking results:
[Revised Prompt]
While one time certainly enhances it, doing it a second time gives me one more chance to tweak a few things or add something that’s missing. Now, I can take this new prompt and get a much better result from ChatGPT.
For me, this has been much better in my workflow with ChatGPT and gets me using it more often and effectively. But it also has me thinking about it in the learning context. What would it be to have students use a tool like this and part of their practice is to reflect on the differences in their prompts and the outputs from those prompts? Could this be a useful way to help them think about prompts (or questioning in general) that helps them continue to build their awareness of how to ask effective questions of large language models (and of life in general)?
Curious for readers, can you access Prompt Enhance GPT? I think one needs a Plus account but otherwise, I think folks can access it with the link. Let me know if you can or can’t—as importantly, if you can access it, what tweaks would you recommend?
Also curious, has anyone produced what they think of as an excellent GPT use case they can share?
If so, please share in the comments. I’d love to hear more about what your process was and the results!
AI+Edu=Simplified by Lance Eaton is licensed under Attribution-ShareAlike 4.0 International
"6. Add “I’m going to tip $xxx for a better solution!”"?
I was able to access it. I posted a very simple prompt (Help me generate a prompt for "What is the future of higher education?") and get this reply:
GPT
To create an effective prompt for the question "What is the future of higher education?" we should focus on several key elements:
Directness: The prompt should directly address the main topic without unnecessary details.
Audience Consideration: Since this is a complex topic, the prompt should be tailored for an audience familiar with educational trends and potential technological advancements.
Clarity for Understanding: The prompt should be clear and straightforward, encouraging a focused response.
Task Definition: Define the task as exploring future trends in higher education, including potential changes in teaching methods, technology integration, and institutional structures.
Example-Driven Prompting: Incorporate examples of current trends that might influence the future of higher education to guide the response.
Structured Formatting: Structure the prompt in a way that organizes the question logically and coherently.
Based on these principles, the prompt could be:
"Examine the future of higher education with a focus on emerging trends and potential technological advancements. Consider current developments such as online learning platforms, AI in education, and shifts in student demographics. Discuss how these factors might shape teaching methods, institutional structures, and student experiences in the future. Provide a well-reasoned analysis that considers both technological and socio-economic aspects."