My First Lesson.

[errant shuttle]

My first lesson is about collecting our experiences in actually solving problems with prompting in ChatGPT.

I want you to write your problems, yours solutions to these problems, and examples of outputs, your flow of inputs to get to a desired Output, The Context, literally anything you can think of or postulate about, expand on it. So in essence my first lesson is our first gathering. 🙂 as I am going to take the time to collect mine and present here over the next few days, I will try to make it in a understandable format.

This is related to ChatGPT the website not the API. though some knowledge, language, or techniques may transfer over to that. this isn't about the API's yet, Thank you for your cooperation.

If someone is posting please let them finish before posting. This will help keep it organized. Thank you for reading this.

it my experience it respond it responds well to structures and in structures.

Here is a example of a programmatic word structure that i made to use in chatGPT.

expand Prompt Engineering
over Natural Language Processing
over Grammar
over Semantics
over use proper Syntax
over be Concise
over be Courteous
expand Machine Learning
over Neural Networks
over Deep Learning
over Natural Language Understanding
expand Real-Time Interactive Systems
expand Creative Prompt Engineering
over Conversational Agents
over ChatGPT
expand Content Generation
over Text Generation
conquer Evaluation and Improvement
over Metrics
over User Feedback
over Fine-tuning
over Transfer Learning
over Bias Detection and Mitigation
over

this is the output that I see.

So what am I doing here that matters?. first I am using the style of a programmer to write normal text.
Second I am using repeated words to structure relations in text. these words should be removed from the output as they repeat so much. next I used 1 single different word that is conquer which is a software design technique called (Divide & Conquer) so Programming 'structure' and wording, with sprinkles of word play. You may also notice the over at the end of my text with no text after it. I use that to close out the response, so it doesn't say anything after producing my desired output.

note the topics and keywords in that are vast topics of things it is expected to variate on those things each time the prompt is used. You may also see the same structure in the output as input. this is one way to structure output of chat GPT.

[errant shuttle]

To be honest about the first Post I use it to understand a topic and the terms involved in it, as i am not a expert in machine learning, i just research it a bit. 🙂

so you can replace the topics and subtopics with something you like and it should show you some stuff that is related to it.

which imho is the coolest part.

Modify, Adapt, Overcome, Live don't exist. Love more and Live more. This is my advice to you.

I am going to cover more output structuring & formatting in my next post. I think this is a common problem and I have attempted it many times with varying degrees of successes and failures.. 🙂 Thank you.

[tired harbor]

@errant shuttleinteresting

can you show a typical case?

[errant shuttle]

explain that a little. please.

the use case for this is topic drilling.

you start with a topic and branch sub topics downwards to add context to chat GPT.

or you build a application with a top down approach to details. 😉