What is a Custom GPT?
Using grammars to constrain llama.cpp output
LLMs Just Don't Understand
The answer is the easy part