AI isn’t far from becoming an integral part of our lifestyle. Numerous chatbots such as Gemini and ChatGPT have already taken over the internet. Almost all modern-day smartphones have also started coming with tons of AI-packed features such as AI chatbots and AI-erasers for images.
However, even though using AI is fairly easy, you might not get a satisfactory response every time. And to use any AI properly, you should know of the prompting techniques. Prompts are a very important factor that defines your point to the AI and even the slightest error in the prompt can lead to mediocre responses.
If you also want to learn about the best prompting techniques, then you are at the right place. Let’s get started.
1. Role Based Prompting
Giving certain roles to the AI before asking any question can help you to get a more defined response. AI, or to say a large language model tries to get a neutral approach to everything, which is good but not always the best. So in some cases, you should know how to use this role-based prompting.
Let’s say I have to ask about the origin of Black Hole from ChatGPT. So in such a case, instead of directly asking the bot to explain to me how Black Holes originated, I can give a prompt that reads “Act like a space scientist and explain to me the origin of Black Holes.”
This prompt will ensure that you get fairly reasonable and easier-to-understand answers, instead of word jargon. Persona prompting also ensures a more human live conversation. You can also test it yourself to see the difference.
2. Shot Based Prompting
This is an important method of giving prompts to the AI. The word “Shot” here refers to “Examples.” One shot means one example, zero-shot means zero example and multi-shot means giving multiple examples. By giving the AI a certain example of what you want, you will get a more detailed and personalized response.
Here’s an example of the prompt written using “One Shot Prompt Technique”-
“Write a custom and short self-help book. You can take an example of Atomic Habits.“
This prompt will give the AI an idea of what exactly you want. In the following prompt, we have used an example of the book called Atomic Habits to give us a short and custom self-help book. If the same prompt had been written using the zero-shot prompt, you would just ask the AI to write you a self-help book without any example. If it had been written using the multi-shot prompt, we would have given it 4-5 different examples.
3. Self-Criticism Prompting
As told earlier, AI or LLMs are designed in such a way that they always try to give a natural answer. But if you are not satisfied with the response, and want a more personalized answer then you can use this prompting technique.
Self-criticism prompting, as the name implies, involves AI giving prompt in such a way that says it to criticize its response and give a better or improved version of its response.
Here’s an example of self-criticism prompting:
“Explain to me the pros and cons of mobile phones. Now, critique the response and provide me with a redefined version based on the given critique.”
4. Negative Prompting Technique
This prompting technique is fairly easy. You just have to defy the AI on what you want and what not. Let’s take the past prompt for example here as well. But if you write it using this technique it would read “Explain to me the pros and cons of mobile phone, but do not exceed 5 points each.
In the following prompt, you have clarified to the AI that you do not want more than 5 points and therefore the response you will get from it won’t have more than 5 points. If you hadn’t given this negative prompt, it might end up giving 10-15 points which makes the response hard to understand and messy for you.
5. Chain of thought prompting
This prompt is similar to the way you ask your maths teacher to explain the steps. This prompt technique involves asking AI to explain the answer or the response step-by-step. This prompt is useful when you are asking AI to solve complex problems or need a definition for points made.
Here’s an example of a chain of thought prompting:
“A bus is running at the speed of 60KMPH and the next two stops are at 82KM and 126KM respectively. The bus starts at 8:00 AM. When will it reach the two stops? Explain the answers step-by-step with reasonings.“
If you don’t give this chain of thought prompting, the AI will directly give you the answer which might leave you confused.
Final Words
These were the top prompting techniques you should know before using AI. Prompts are an important part when it comes to using AI tools and hence knowing about it and the best technique for it will make the use of AI not only easier but better for you. You can also test all the prompts and we are pretty sure that you will be able to get the best response or at least better from the original response.
Related Articles:
- How to Create Custom Windows 11 Screensaver [2 Methods]
- How to Activate AI-powered Copy Paste Feature on Windows 11
- OpenAI launched ChatGPT for Windows users, but there’s a catch
- How to Use Flux AI Locally on Your Computer For Free
- How to Run and Use AI Locally on Your Low RAM PC
- How to Use Flux AI Image Generator for Free
