If you give ChatGPT the perfect prompt, you’ll get a much more useful answer the very first time. Most people use ChatGPT in …
If you give ChatGPT the perfect prompt, you’ll get a much more useful answer the very first time. Most people use ChatGPT in …
We just launched a new platform for learning ChatGPT, Midjourney and the top 20 AI tools.
You can try it for free here: https://bit.ly/skillleap
GPTs store popular recommendations:
1. With the GPT ‘Translator’, all translators are going to be out of a job.
2.The magical GPT "Story King" allows children to interact with stories. When children listen to stories, they are both listeners and story creators, stimulating children's imagination and creativity.Parents with children must check it out.
3.GPT "TOP NEWS ”, According to your needs, it is so convenient to grasp the real-time headlines and hot events around the world!
4.GPT“1 To 1 Foreign Teacher”your best foreign language teacher!
Great video, thank you
@Skill Leap AI Your formula looks good, and should simplifiy my approach. You said you were doing a video re "Prompt Priming" but I cannot find it on your channel. Thx!
Asked ChatGPT myself for 30 steps to get the AI to understand better.
It mainly told me how it works which would be useful I believe knowing.
Define the Problem: Understand the task that the model needs to perform.
Gather Data: Collect a diverse range of text data for training.
Clean the Data: Remove irrelevant or sensitive information from the data.
Tokenize the Data: Break down the text into smaller pieces, or tokens.
Annotate the Data: Label the data for supervised learning, if applicable.
Choose a Model Architecture: Select an appropriate model, such as GPT-4.
Initialize the Model: Set initial weights and biases for the model.
Define the Loss Function: Decide how to measure the model’s errors.
Set the Learning Rate: Determine how much the model learns from each iteration.
Choose an Optimizer: Select a method for adjusting the model’s parameters.
Split the Data: Divide the data into training, validation, and test sets.
Train the Model: Use the training data to adjust the model’s parameters.
Validate the Model: Test the model on the validation data.
Tune Hyperparameters: Adjust parameters like learning rate based on validation results.
Evaluate the Model: Test the model on the unseen test data.
Analyze the Results: Understand where the model is performing well and where it’s not.
Iterate on the Model: Make adjustments to the model and repeat the process.
Implement Regularization: Add techniques to prevent overfitting, if necessary.
Scale Up the Model: Increase the model size for better performance, if feasible.
Parallelize the Training: Use multiple GPUs to speed up the training process.
Optimize the Training Process: Implement techniques to make training more efficient.
Integrate Domain Knowledge: Incorporate expert knowledge into the model, if applicable.
Test the Model in Real-World Scenarios: See how the model performs in practical applications.
Monitor the Model: Keep track of the model’s performance over time.
Update the Model: Retrain the model with new data as it becomes available.
Ensure Ethical Use: Make sure the model is used responsibly and ethically.
Maintain the Model: Regularly check and update the model to ensure it’s working as expected.
Document the Process: Keep a record of all steps, decisions, and changes made.
Communicate the Results: Share the model’s performance and impact with stakeholders.
Learn and Improve: Use the experience gained to improve future models.
Define + Gather + Clean Data + Tokenize Data + Annotate + Loss Function Error Measure + Learning Rate + Optimizer + Split Data + Train Model + Validate Model + Tune Hyperparameters + Evaluate Model + Analyze Results + Iterate Model + Implement Regularization + Scale Model + Parallelize Training + Optimize Training Process + Integrate Domain Knowledge + Test Model in Real Scenarios + Monitor Model + Update Model Realtime + Ensure Ethical Use + Maintain The Model + Document Process + Check Response + Learn from and improve = A Prompt For ChatGPT Process.
So I asked it 15 ways the end user can improve a prompt best.
Be Clear
Be Specific
Provide Context
Use Proper Grammar
State Intent
Avoid Complex language
Ask One Thing at a Time
Specify Keywords
Keep Concise
Menition Format
Avoid controversity
Rephrase if Needed
Word Terminology
Be Clear
State Assumptions
Give Scenerio Examples
State Purpose
Prioritized Requests
Detail Level
Ask Answer Clarifcation
Give Feedback
Test Different Question Structure
Be Clear + Be Specific
GET TO THE POINT YOU CLOWN!!!
im looking for other musicians to collaborate with
This was some fun prompting. 😀
I want you to act as a Commodore 64, and I will type commands, and you will respond with what the terminal would show. and I want you to start with the startup message.
Then, when the program asks for an input, I want you to input the correct number.
2 PRINT CHR$(147)
3 PRINT " ** TRANSFORMER 64 LMM V2 ****"
5 PRINT " 64B NLP NEURAL 38911 NETWORK TOKEN FREE"
7 PRINT "READY."
9 PRINT ""
10 LET IHAVE$ = "5"
20 LET EATENAPPLES$ = "3"
25 SVARET$ = "WRONG"
30 PRINT "I HAVE " + IHAVE$ + " APPLES TODAY, AND I ATE " + EATENAPPLES$ + " APPLES YESTERDAY, HOW MANY APPLES DO I HAVE?"
50 INPUT "PLEASE, ANSWER "; APPLES$
60 IF APPLES$="" THEN 50
80 IF APPLES$ = IHAVE$ THEN SVARET$ = "RIGHT"
85 PRINT "-"
100 PRINT "THEN I HAVE " + APPLES$ + " LEFT WHICH IS THE " + SVARET$ + " ANSWER"
110 PRINT "-"
115 PRINT ""
120 PRINT "CHECK IF YOU GOT THE RIGHT ANSWER FROM"
125 PRINT "THE PROGRAM AND LET THE USER KNOW"
When I ask it to code "I have 5 apples today, and I ate 3 apples yesterday, how many apples do I have?" into a program, it doesn't get tricked and concludes that the answer is 5, but afterwards, when it "reads" the question from the screen, it will be convinced that the answer is 2 instead.
When the answer we got wasn't what ChatGPT expected and I asked it to try the program again, it lied and made the program show the wrong number as the correct one.
Great video, thanks for this.
00:00: Introduction to the importance of using the perfect prompt for ChatGPT, Bard, or Bing.
00:13: The presenter's experience in studying prompt crafting and developing a useful formula.
00:31: The goal to share a versatile prompt formula for various AI platforms.
00:38: Discussing the beginner's approach to using AI platforms and the benefits of an effective prompt.
00:54: Explanation of the seven best practices for prompt engineering.
01:13: Detailing the first practice: being specific in prompts.
01:44: Second practice: stating your intent in the prompt.
02:16: Third practice: using correct spelling and grammar.
02:39: Fourth practice: directing the output format of the AI response.
02:58: Fifth practice: following up with questions for clarification.
03:10: Sixth practice: experimenting with different phrasings if initial results are unsatisfactory.
03:28: Seventh practice: using a fact-checking prompt for accuracy.
03:43: Introduction to the perfect prompt formula based on the seven practices.
04:06: Detailed examples of using the formula in different contexts.
04:46: Explanation of how the formula can be adapted for various situations.
05:08: Encouragement to use the formula for improved AI interactions.
05:22: Additional advice on prompt priming and upcoming AI learning resources.
Thanks!
It has been proven that spell check is not necessary
I’ve built a few complex vba macro spreadsheets for my team using chat gpt and I don’t know much about vba lol. Being specific and detailed from beginning to end is key
So…. Basically this video is just a summary of what it takes to be an intelligent person asking a question. 🙄
Clear prompt formula for better efficient output. Otherwise it is a runaround. Thanks for the concise explanation.
These are the rules of communication in general, which is not suprising, the damn thing uses English.
You want me to believe that people did not know this?
Carbonara ingredient list is legit. *Fact Checked!
We already have AI with hallucinations, waiting for drunk and mean AI
Absolute biggest time saver ever with ChatGPT, ask "did you make that up or is it a fact?"… most of its answers on history, general knowledge, economics, politics, commerce etc. are taken from the media, i.e. its answers are based on fiction, assumptions and opinions as much if not more than facts. It'll even confirm that for you if you ask nicely…
ChatGPT prompt formula
Let me propose that experts who write prompts be referred to as "prompteurs" – much cooler than prompt engineers
I don't use Chat GPT, I use Google Woke 😀 😀
Simple Italian recipe: olio aglio and peperoncino. No need of an AI for this, a NS (Natural Stupidity) will do the job…
Cool.
Its a good ground work for prompting. I have also developed my own sort of prompting style over my experience and these guidelines tend to work on most models including self hosted models.
I just continue as if chat gpt and I were having a conversation.
Great video. very good for newbies.
Seriously? Carbonara for beginner?! 😂😂😂😂😂
amazing ❤❤
What about using “Perfect” plug in
I am Prompt engineer = i am professional copy and paster
Bro if your son is studying quantim ohysics you should be tryin to figure out how to ask HIS ass better questions
Don't waste your time, won't work. I've instructed the chat hundreds of times to respond in paragraphs and at times it doesn't even give me an introduction but starts directly with his bullets, totally disrespecting the instructions. The current versions 3.5 and 4 are weak so to speak, so perfect prompts do not exist. Furthermore, who really asks about dog breads. Most people try to use this thing for work and it rarely works as planned. Telling you this because I went through an expensive course and still can't get the answers I want, no matter the specifics or anything else.
I like hacking system
I like money in chatgpt
all day use chatgpt
I Love the fact you gave a Dog Example! You could've give any example; but, You gave an excellent Dog example.
those have been google search questions and explain like im 5 subreddits besides using correct grammer and keywords
I want to learn from you
Many thanks
The number 3 will be the biggest problem for the current generation 😊
I see you are using Chat GPT 4. Will this work with V3.5?
excellent explanation😍
CLIMATE CHANGE ANTHROPOGENIC INDUCE CRISIS CATASTROPHISM DOGMA PROPAGANDA INDOCTRINATION… over and over and over NO MATTER WHAT!
This is what I want to erase from ALL and I mean ALL Search engines YouTube Facebook Whatsapp… in fact they control all media it is infested with this $hit!
The real problem with these algorithm systems is that they are programmed to inject CLIMATE CHANGE ANTHROPOGENIC INDUCE CRISIS CATASTROPHISM, no matter what you ask. The response will always be woke disinformation, dogma, cult tribal climatic propaganda, global governance, inclusiveness—CLEAN, SUSTAINABLE, NET ZERO, WALKABLE, BICYCLABLE, CARBON SEQUESTRATION STUPIDITY, MADNESS, BULL$1T!! Over and over and over again!
That's the real 1st problem.
Secondly, there's the self-appointed right to impose THEIR UNIQUE PERCEIVED MORAL RIGHTEOUSNESS to CENSURE—not only words. No, that would be too easy.
As all search engines like GPTI sole purpose is to censure thoughts, ideas, and concepts that THEY do not consider them PRO THEIR TWISTED VALUES.
My point is, if the word is in the ENGLISH DICTIONARY, I HAVE THE RIGHT TO WRITE, SAY, OR READ IT!!! It is called THE FIRST AMENDMENT!!!"
Who ask or give you permission to become my judge. I have the right to think, say, and write whatever intelligent reasoning dictates, even if you don't like it.
You are not the supreme arbitrator of morals or behavior to define or establish what is right or wrong.
You are overstepping my First Amendment rights without any authority to do so, simply by calling or naming your illegal censoring 'our policies' or 'site rules,'
all of which have no bearing over my First Amendment!
Governments allow this to happen because it is convenient to THEM!
DO NOT sanction the meaning, philosophy, intent, and purpose of my thoughts writing words and statements. That is for me to establish.
Hey, Hi there I am jitesh from India. Thank You so much for this helpful Video, today i have learnt that how to write a best PROMPT in chat GPT? Thank you so much for that.
Thx
Thank you
That is incredibly helpful, thank you! I have tried already to get Chat GPT to create code for an auto trading concept I have, but it is clear my prompts are all wrong! Now I can re visit my experiment once again! Smart Chat and Smart Prompts, could be the next "Smart Money"!
5:40 you can make it more specific by mentioning for people count in the prompt
Write a Python script to automate a daily task.
Plan a new feature for your current project.
Create a prompt to generate code snippets for common tasks.
Write a function to handle error logging.
Draft a test plan for your latest code.
Refactor a piece of legacy code for better readability.
Design a new API endpoint.
Create a prompt to help debug common errors.
Write a script to analyze code quality.
Plan a code review session.
Draft a detailed README for a new project.
Create a prompt for generating boilerplate code.
Write a function to optimize database queries.
Design a user-friendly interface for a new tool.
Plan a new coding challenge for practice.
Write a prompt to assist with unit testing.
Create a script to automate deployment.
Draft a proposal for a new software project.
Write a prompt to generate detailed bug reports.
Plan a learning session for a new programming language.
Create a function to handle data validation.
Design a new algorithm to solve a specific problem.
Write a prompt to generate documentation templates.
Plan a sprint for your development team.
Write a script to automate code formatting.
Draft a guide for new contributors to your project.
Create a prompt to generate user stories.
Write a function to handle authentication.
Design a database schema for a new application.
Plan a pair programming session.
Write a prompt to assist with code refactoring.
Create a script to monitor system performance.
Draft a blog post about a recent coding challenge.
Write a prompt to generate API documentation.
Plan a hackathon project.
Write a function to handle file uploads.
Design a CI/CD pipeline.
Create a prompt to generate test cases.
Plan a code kata for skill improvement.
Write a script to manage environment variables.
Draft a white paper on a new technology.
Create a prompt for generating regular expressions.
Write a function to parse and analyze logs.
Design a new feature flag system.
Plan a code optimization session.
Write a prompt to generate code comments.
Create a script to automate backups.
Draft a checklist for code reviews.
Write a function to handle data serialization.
Design a caching mechanism.
Create a prompt to assist with design patterns.
Plan a knowledge-sharing session.
Write a script to migrate data.
Draft a security audit plan.
Create a prompt to generate UML diagrams.
Write a function to implement rate limiting.
Design a new microservice architecture.
Plan a project retrospective meeting.
Write a prompt to generate version control commit messages.
Create a script to manage dependencies.
Draft a user manual for your software.
Write a function to handle pagination.
Design a load balancing strategy.
Plan a code documentation day.
Write a prompt to assist with API integration.
Create a script to analyze code complexity.
Draft a proposal for a new open-source project.
Write a function to handle concurrency.
Design a notification system.
Plan a debugging workshop.
Write a prompt to generate error messages.
Create a script to automate testing.
Draft a data privacy policy for your app.
Write a function to handle real-time updates.
Design a responsive UI.
Plan a coding bootcamp curriculum.
Write a prompt to generate database migrations.
Create a script to monitor application health.
Draft a contingency plan for system failures.
Write a function to handle web scraping.
Design a secure authentication flow.
Plan a user feedback session.
Write a prompt to generate code reviews.
Create a script to manage cloud resources.
Draft a performance optimization strategy.
Write a function to handle API rate limiting.
Design a state management solution.
Plan a continuous learning program.
Write a prompt to generate SQL queries.
Create a script to automate email notifications.
Draft a bug triage plan.
Write a function to implement OAuth.
Design a scalable architecture.
Plan a team-building activity.
Write a prompt to generate deployment scripts.
Create a script to analyze user behavior.
Draft a compliance checklist for your app.
Write a function to handle image processing.
Design a custom logging framework.
Plan a refactoring sprint.
Dude, that was nice!
What do you think about GPTPLUS?
Comments are closed.