In a world where artificial intelligence (AI) is increasingly becoming part of our everyday lives, new forms of interaction between humans and machines are emerging. Large Language Models (LLMs) such as ChatGPT, Microsoft Copilot, Google Bard and others are at the forefront of this development and offer unprecedented opportunities to collect, learn and communicate information. But with great power comes great responsibility – especially in the way we interact with these models. This is where “prompting”, i.e. asking questions and giving instructions to an LLM, plays a decisive role.

Why is correct prompting becoming increasingly important? The answer lies in the nature of LLMs themselves. These models are trained to react to the user’s input – they are like mirrors that reflect our queries and generate answers based on them. A precisely worded prompt can make the difference between a general, unhelpful answer and a profound, tailored insight. In an age where information can be flooding and sometimes overwhelming, good prompting allows you to unleash the true potential of LLMs – whether in academia, business or for personal use.

(Prompt generated by Sophie Hundertmark, executed by Microsoft Copilot in Google Chrome 01-2024)

In this article, I will give you some important best practices for dealing with LLMs and correct prompting.

What is a Large Language Model (LLM)?

A Large Language Model (LLM) like ChatGPT is basically a very advanced computer program that has been trained to understand and generate human language. Here are some simple points to explain it:

1. Training with large amounts of text: The model learns by analyzing huge amounts of text (such as books, articles, websites). As a result, it understands how language is normally used.
2. Pattern recognition: It recognizes patterns in language, how words are usually combined to form sentences and paragraphs. It also learns to extract and interpret information from texts.
3. Answers like a human: Based on what it has learned, the model can answer questions, write texts or hold discussions just like a human would.
4. Adaptability: It can adapt to different topics, language styles and requirements depending on how it is asked.
5. Not perfect: Although it is very advanced, the model is not perfect. It can make mistakes or sometimes misinterpret information because it is still relying on the patterns and examples it has seen during training.

In short, a Large Language Model is like a very comprehensive, automated reference book that has learned to communicate and respond in a human-like way.

(Prompt generated by Sophie Hundertmark, executed by ChatGPT 01-2024)

You can read more about ChatGPT in my ChatGPT explanation article.

Prompt Guidelines

Over the past few months, I’ve studied a variety of academic papers, blog posts and YouTube videos to get new tips for better prompting. I then tested and optimized these tips myself and summarized them in 26 Prompt Guidelines.

You can buy and download the guidelines directly by clicking on the button below.

Further useful tips

I also offer workshops, coaching and starter courses around ChatGPT and LLMs.

And of course you can always contact me directly if you have any questions. Just send me a message via WhatsApp or e-mail.

Or come directly to my WhatsApp group – where I regularly post use cases, news, best practices, events and much more about chatbots, ChatGPT and co.