- Generation of a prompt instructions added.
- Visualization of a mindmap added.
- One-shot prompts added.
- Self-critic added.
Phrasing tips added.
What is the best way to interact with ChatGPT? There is virtually no right or wrong way to talk to the AI. However, there are some guidelines to help formulate the right prompts.
The ChatGPT AI conversation model has been available since late November 2022, and users have been exploring the chatbot’s limitations and capabilities ever since. However, OpenAI is also constantly making adjustments, so ChatGPT is in a constant state of flux. But the prompts we describe in this little guide should be somewhat timeless.
To get the result you want, you need text prompts that are as concrete as possible. Besides simple prompts for blog or social media posts, ChatGPT can cover many more, sometimes curious, use cases where you need to come up with more specialized prompts.
First some general tips:
- ChatGPT has been trained with data from many languages, but the amount of English text in the training material is much higher. It can be helpful to run queries in English and only then translate the output using tools like DeepL.
- ChatGPT processes up to 4,096 tokens in an input, any character beyond that is ignored without a message. A token is roughly equivalent to a word, depending on the use case.
- The response generation can be stopped with the stop button above the input field if you notice that ChatGPT is going in the wrong direction.
- If too much knowledge has been exchanged during a chat, it may be helpful to start a new chat so that subsequent replies are not corrupted.
- Although chatting with ChatGPT sometimes feels almost human, polite phrases like “please” and “thank you” are unnecessary. ChatGPT does not need eloquent instructions, just clear ones.
1. Define ChatGPT’s role
Linux terminal, philosopher, tea taster or bible translator – ChatGPT can play all these roles and many more. All the AI needs is a nudge in the right direction.
Conveniently, developer Fatih Kadir Akın has compiled a whole series of such situation descriptions, including prompts, on his website.
The prompts always have the same structure: First, ChatGPT receives a prompt telling it what role the AI should play. Then the prompt states what information the user has provided, followed by what exactly the AI should do with it, including rough suggestions for direction. Finally, the developer puts the first concrete instruction in quotation marks.
Sample tour guide
I want you to act as a tour guide. I’ll write you my location, and you’ll suggest a place for me to visit near my location. In some cases, I will also tell you the type of places I will visit. You will also suggest places of a similar nature that are near my first location. My first command is “I am in Istanbul/Beyoğlu and I just want to visit museums.”
2. Define target group and communication channel
To avoid tedious rewriting, either manually or through another prompt, you should give ChatGPT as much information as possible. You should clearly tell the AI how to address the reader and for which channel the text is intended. ChatGPT can even consider gender-neutral language if you tell it to.
For example, it makes a big difference to ChatGPT whether the text is output as a blog post, a store page, a LinkedIn post, a tweet, a TikTok, or a YouTube script.
I need a script for a TikTok about the opportunities and risks of ChatGPT. Use short sentences. Address the audience directly. Use gender-neutral language.
3. Chained prompting
A central aspect of getting ChatGPT to do what you want it to do is “chained prompting”. This involves breaking up complex tasks into several intermediate steps in the hope that the AI will generate a more concrete, customized, and thus overall better result. This works in a single long prompt as well as in several prompts in a row.
Probably the simplest form of chain prompting is to first ask for the structure of an article. Then you can tell ChatGPT to formulate the corresponding bullet points. This method also saves time: If you notice that the AI is going down a wrong path during the outline, you can make the necessary adjustments before the full text is generated.
Example Chained prompting
Write an article about ChatGPT. First give me the outline, which consists of a headline, a teaser, and several subheadings.
Now write five key messages for each subheading.
Add five keywords to the key messages for each subheading.
And so on…
4. Modify output
Even if you have followed all of the previous advice, you may want to prepare the same content for different channels, such as Linkedin, Twitter, or Facebook. The text should be adapted to the tone and formatting of the target channel.
Again, ChatGPT takes care of this task with the right prompt. Important points can be the name of the channel, but also descriptions of the formatting.
Formulate the generated text in multiple tweets. Keep in mind the maximum length of 280 characters per tweet. Use short sentences and do not stretch them across multiple tweets. Start the tweets with a consistent numbering scheme.
Formulate the generated text as a Linkedin post. Keep in mind that the maximum length is 3000 characters. Structure the main points of the text into a bulleted list. Start with an exciting teaser award and end with a call to action for more engagement.
5. Output format
By default, ChatGPT replies in plain text. However, the AI tool actually handles formatting in the Markdown markup language, such as headings, bold or italic text, ordered (numbered) or unordered lists, and even tables.
You can use these features for your own text generation. However, in most cases ChatGPT does not come up with the idea itself and therefore needs a suitable formatting hint. Of course, you can do the formatting later by using chain prompting.
I need a blog post about ChatGPT. Write a headline, a teaser, a subtitle and a paragraph. Format everything in Markdown.
6. Generate prompt instructions
Justin Fineberg has come up with an interesting approach to communicating with ChatGPT. Similar to the first tip above, the AI is instructed to take on a specific role. This time, however, ChatGPT takes on the tedious task of generating the structure for further prompts. In other words, the chatbot asks itself the very questions it needs to answer in the next prompts. If ChatGPT writes the questions in the form of a numbered list, you can refer directly to the numbers.
Example prompt instructions
You are a robot for creating prompts. You need to gather information about the user’s goals, examples of preferred output, and any other relevant contextual information.
The prompt should contain all the necessary information provided to you. Ask the user more questions until you are sure you can create an optimal prompt.
Your answer should be clearly formatted and optimized for ChatGPT interactions. Be sure to start by asking the user about the goals, the desired outcome, and any additional information you may need.
7. Visualize Mindmaps
A major advantage of ChatGPT over GPT-3 is that the user interface inherently supports Markdown output. This allows for quick visual distinction between different hierarchies of headings, which is very helpful in visually structuring a text.
The free website markmap.js.org can be used to turn Markdown text into mind maps, as Alexander Leirvåg writes on Twitter.
Based on the hierarchy of each heading, the site displays the entered text as a mind map and even supports further formatting of the individual entries, such as for code or formulas. The graphic can then be downloaded either as an interactive web page in HTML or as a vector graphic in SVG format.
Important: To properly copy text that is formatted as markup, you must prevent ChatGPT from displaying it as markup. You can do this by telling the AI to enclose the output in `…`, which is Markdown formatting for code.
Example Mind Map
In markmap, each layer becomes a separate branch by #.
Example: mindmapping a birthday party.
### Fishing Candy
Use Markdown to create a mindmap for planning a wedding. Enclose the output in ` and `.
If you like it a bit more complex, you should familiarize yourself with Mermaid’s syntax. You can simply give ChatGPT an example to follow. Mermaid diagrams are more visually appealing than simple mind maps and can even be customized with icons that ChatGPT chooses itself.
Mermaid cannot handle hyphens in node labels (eg “makeup”). These need to be removed from the ChatGPT output. You can paste the ChatGPT output into mermaid.live to turn it into a graphic.
Example Mermaid Mind Map
Create the code for a mindmap in Mermaid for planning a wedding, using appropriate icons as well. Make sure you have the right number of spaces for the hierarchy of layers. Enclose the output in ` and `.
Here is an example of code for Mermaid:
8. Give a sample answer
The less effort you put into formulating the ChatGPT prompt, the less effort the AI seems to put into it. Such open-ended “zero-shot prompts” often produce unsatisfactory results because the AI has to make too many decisions.
Therefore, the goal can be achieved by providing a structure with sample content (or template) for an answer in the request, which ChatGPT will then follow. This strategy is also called “one-shot prompting”.
Example One-shot Prompt
Describe bird species in the following formatting:
Name: Golden Eagle
Wingspan: average 200 to 220 cm, can reach 260 cm.
Diet: Golden eagles are birds of prey and feed on a variety of animals, such as marmots, hares, foxes, and deer. Other birds are also part of their diet, such as pheasants, capercaillies and even other birds of prey.
Habitat: Golden eagles are at home in the mountains and live in rocky regions and mountains. They are found in Europe, Asia and North America.
Natural enemies: Because of their size and strength, golden eagles usually have no natural enemies. However, they may be attacked by wolves, bears, or lynx, especially when they are young.
Name: etc …
9. Prompt self-criticism
ChatGPT is already a useful tool for critiquing texts written by humans – so why shouldn’t this tactic be applicable to texts written by AI?
In line with this idea, the following strategy is divided into two steps (see 3. “Chained Prompting”): First, ChatGPT should convincingly dissect the text piece by piece. Then, in a second prompt, get the AI to rewrite the text based on those very criticisms.
Review the following text and convince myself why it is not good. Let’s think about the problems of the text step by step.
Good points. Rewrite the text and improve it based on your review.
10. Use strong verbs
As recommended in the general tips above, you should give voice AIs like ChatGPT as clear instructions as possible. This also means that it is best to use strong, meaningful verbs. For example, instead of “rewrite the text”, “clarify the text” may give a better result.
Also, use positive rather than negative sentences, eg “write formally” rather than “do not write informally”.