Your question is Prompt Length vs Context Budget. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
You are designing prompts for an LLM feature that regularly works with long inputs, such as long documents, chat histories, or many retrieved passages. You have noticed that adding more instructions, examples, and context can improve answer quality in some cases, but it also increases token use and can hurt focus.
How do you balance prompt length and token budget with performance when designing prompts for long-context windows?