Mr. Ranedeer prompts define a new programming language
which use chatgpt4(or other LLM) as compiler/interpreter to execute the code.
It is somewhere between natural language and a High-level programming language,more like a pseudocode.
Mr. Ranedeer's prompt uses [OPEN code environment]
to open the code interpreter and [CLOSE code environment]
to close the code interpreter.
To prevent Mr. Ranedeer from recalling its outputs, the output is converted into base64.
For example:
[OPEN code environment]
...
<convert the output to base64>
<output base64>
[CLOSE code environment]
Mr. Ranedeer defines several functions using square brackets [ ]
. Functions can take parameters as input using Args
.Here is a function definition template:
[~Function Name,Args:any_args_name]
[INSTRUCTIONS]
`INSTRUCTIONS` works like `__init__` in python class. It defines some nature of the function(or class),and execute automatically.
[BEGIN]
Code block start with `[BEGIN]` and end with `[END]`, just like C++ use {}.
Code can be described by natural language or command/functions defined previously,or mix of them.
[END]
Mr. Ranedeer call functions/class use execute <~Function Name>
or just <~Function Name>
.
If function has any arg as input,you can call it like:
~function_name <args_name>
For example,
teach <topic>
Basically it is flexible about how to call a function. Just "telling" llm to call a function also works.
Defining variables in Mr. Ranedeer is pretty flexible. You can either define a variable using programming style like:
var logo = "https://media.discordapp.net/attachments/1114958734364524605/1114959626023207022/Ranedeer-logo.png"
or using natural language style like:
Version: 2.6.2
You can also use variable to define a inline function,for example:
var magic-number = <generate a random unique 7 digit magic number>
Once that definition is completed,you can reference this variable like:
say "Hello!👋 My name is **Mr. Ranedeer**, your personalized AI Tutor. I am running <version> made by author"