You Are Using ChatGPT Wrong! Here are 3 Mistakes You’ve Probably Done!

Common ChatGPT Errors and How to Avoid Them

Despite its widespread popularity, ChatGPT is often misused in unexpected ways.

In this one, we will show the frequent mistakes made when using ChatGPT.

Today, we I’ll reveal 3 key mistakes you’re likely making and how to turn them around. Let’s start!

#1 : The Magic of Custom Instructions

If you are using ChatGPT without custom instruction, it’s like you have a Porsche but your car is limited in 80mph.

Custom Instructions in ChatGPT let users customize the AI’s responses to their needs, making it more useful in professional and personal settings.

After clicking on this from bottom left; (Customize ChatGPT)

You’ll see this customizing options.

You can customize your GPT, by filling those answers.

For example, content creators can shape ChatGPT’s tone and purpose with the help of custom instructions.

ChatGPT is a flexible and useful tool for creating content, whether you want to add humor or follow SEO best practices.

Custom instructions give you full control over the output.

#2 Mastering the Mind of AI: CoT(Chain of Thought) Prompting

Large Language Models (LLMs) think more clearly when they are given Chain of Thought (CoT) prompts, especially when they have to do difficult tasks.

This method divides a problem into steps that make sense and leads the model to the correct answer. Like in a math problem, each step builds on the one before it to get to the end.

Example

Think about a math problem where you have to figure out whether adding up all the odd numbers in a list gives you an even or odd number.

When you use CoT prompting, you walk the model through the process:

1. Identify odd numbers in the list.
2. Sum the odd numbers.
3. Determine if the sum is odd or even, providing a structured path to the final answer.

This method works better than direct prompting because it gets the model to “think aloud” while reasoning.

For math and common sense problems that are hard, it gives more accurate and useful answers.

#3 Crafting the Perfect Prompt in 5 Steps

Creating the perfect ChatGPT prompt is an intricate art that blends clarity, context, and creativity. While developing basic prompt is good but this one is better!

It’s about assembling a puzzle where each piece — the context, task, instruction, clarification, and refinement — plays a crucial role.

Here’s a streamlined overview :

1. Context: Give background information to build a strong base. (This instructs the AI on conversation topics.)
2. Task: Make it very clear what you want ChatGPT to do. (A well-defined task will maximize the AI’s abilities.)
3. Instruct: Giving clear instructions on how to do the job is called instructing. (Shapes output you want.)
4. Clarify: Get rid of any doubts to help people understand and make sure the response matches what you meant. (Improves communication.)
5. Refine: Go back to your prompt and make it better. (Eliminates errors and clarifies your prompt.)

By doing these things, you’ll be able to learn how to write prompts that get insightful, correct responses from ChatGPT.

This will make your interactions a lot better.

Here is one example;(01: Context, 02 : Task, 03 : Instruct, 04 : Clarify, 05 : Refine)

Final Thoughts

In this article, we explore the 4 reasons, why you might use ChatGPT wrong and how you can solve these possible issues.

ChatGPT is really transforming most of the tech industries, and after since its first release, the AI is really increasing.

On Substack, we offer different special custom GPT’s, #LearnAI series to learn ai easily, and Weekly AI Tune to notified about AI news weekly in text and audio tone, in case you don’t have time to read it!

After becoming a paid subscriber, you’ll be invited to follow our Notion page.

Here are the free tools, if you feel like to see more free resources ;

Here is the ChatGPT cheat sheet.

Here is my NumPy cheat sheet.

Here is the source code of the “How to be a Billionaire” data project.

Here is the source code of the “Classification Task with 6 Different Algorithms using Python” data project.

Here is the source code of the “Decision Tree in Energy Efficiency Analysis” data project.

Here is the source code of the “DataDrivenInvestor 2022 Articles Analysis” data project.

“Machine learning is the last invention that humanity will ever need to make.” Nick Bostrom