Skip to content

Task specific engineered prompts for ChatGPT and other LLM's

License

Notifications You must be signed in to change notification settings

Alex-zun/tot-gpt-promptruly

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

15 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MARKDOWN INSTRUCTIVE MODELS

⚠️ DISCLAIMER

I'm not a data or computer scientist, neither the best writer nor graphic designer, so take my approach with a grain of salt, a fun experiment.

👋 WELCOME!

To task specific engineered instructives for ChatGPT and other LLM's. I'll use this space for posting Ai related prompt Instructive, and templates of my creation.

🤔 MOTIVES

I have only two motives for doing this:

  1. Exploring and experimenting with ChatGPT capabilities in creative endeavours like writing novels, essays, scripts, plays, etc. You get the idea.
  2. Because it's fun!, and I feel like sharing.

💭 CONTEXT

In my excitement, experimentation with ChatGPT on my creative endeavors. I often found myself reading, watching, and testing too many prompts that could fulfil my needs; sadly with little or no concrete results. I find these allegedly “Prompt Engineers” lacking of something fundamental, an 'Oomph' or 'je ne sais quoi'. I find GPT and LLM models in general flexible, well crafted smart tools and, as any tool, used or guided in a certain way, makes the difference, exponentially communicating or doing unexpected things. Numerous prompts out there are overly hyped, too simplistic, not even achieving more than two paragraphs of text, these prompts are just that prompts. There is no indication, none that I'm aware of, that could guide GPT to write or do certain tasks.

As a graphic designer, communicating is fundamental, the process on doing so is essential. ChatGPT assesses that very well in my opinion. Through some research, iterations and testing, I have composed a prompt engineering approach, I call for now “Instructive Model”.

📝 MARKDOWN INSTRUCTIVE MODEL:

The “Instructive model”, is my input approach of prompt engineering for Large Language Models (LLM's) like ChatGPT, OpenAi Playground or any other LLM model. These are written in Markdown files, crafted to Role-play a certain character with well-tuned guides on how to perform by only completing a certain task using Tree of thought Framework in several aspects of user interaction, outputs, and overall experience.

  • I hope my approach squishes GPT capacities, at least, a little bit.
  • Feel free to suggest changes, improvements, and fun stuff to try.

🤖 TEST

In this repository, you find two things

  1. NOVEL-GPT - Storywriter model v0.1 - this is a tested model, i build up for you to test. You can use it as if, or change watherver parameter you like.
  2. A Template - Mold it to suit your specific needs, oho! It's a thing fo beauty, innit?
  3. Test-Transcript - This is the text transcript of one session.

❓ QUESTIONS & RESOURCES

Releases

No releases published

Packages

No packages published