What is actually sent to the LLM ?

[ltrujilloEquancy]

Hello,

I'm using gpt pilot to create a basic CRUD app. Soon, with a gpt-4 LLM I reached the token limitation. Obviously I got over this by increasing the number of tokens that can be fed to the model.

However, I am still wondering what exactly is sent to the LLM at each new task.

Does someone has an idea of how are made the prompts ? Or is there a log file where the prompts are saved so the user can check ?

Thank you in advance for any information about this topic.