If you came to grab my template (ACE) (HFS) for all good LLM's
You can find that on Github here
What is Adaptive Capacity Elicitation?
This is my Cognitive Enhancer originally developed to build Agents with chatGPT4.
Full Research Study: view it in Figma here
GPT4 Agent: use it on chatGPT here
You can use the base template for any task as a base prompt, in combination with whatever task you set for the LLM. If you are unsure, simply copy the text from one of the base files, then add your instructions at the end.
By asking the LLM to use ACE, it will go through an undulating recursive thought process. Depending on your task, your mileage may vary.
There is a basic example for customizing the template for creating image prompts.
When i published this template, i also created the introduction video guide:
The Holographic Filesystem:
This is a simulation of a filesystem, it is not real. The AI is projecting what it believes a filesystem is and how it should work. Furthermore it can hold the information inside those files too, this allows a way to organise the rambling nature of the AI as it iterates through various tasks as it completes whatever instruction is it working on. In this template it is using simple .txt files, it name and create them on it own and is then able to write to them and read from them in later tasks. Because it just creates all of this on it's own, it easy for it to trip over itself.
Using the holographic filesystem, the thoughts are now organised and silent. If an error occurs it can be isolated to a single file and removed, which cleans up cognition for later tasks.
it mimics a linux server CLI environment, and understand the context of instructions, so you can converse with NLP or use raw CLI instructions to work with the LLM. it's possible to add script to certain files you create and allow it to simulate the execution, and do so with accuracy despite the occasional reminder this is all fake and not real.
I will be adding more about this in future videos on the subject.
~~ Part 2 TBA