What is the primary issue when asking an LLM to write code without specific instructions?	The LLM often generates excessive output, including explanations and comments, rather than just the concise code.
What technique can be used to reduce the amount of generated code from an LLM?	Using a carefully crafted system prompt.
What two key elements should a system prompt contain to encourage concise output?	Assigning a specific role and instructing the model to respond as concisely as possible.
How does assigning a role in a system prompt influence the LLM's output?	It guides the model's persona and constraints, leading to more focused and specialized responses.
Provide an example of a system prompt designed to enforce concise Python code.	"You are a Python engineer who writes very concise code."