CCDV-F practice › Model Selection & Optimization · Question 49 of 92
Model Selection & Optimization
A long prompt has several stable sections and one volatile section in the middle. What technique places reuse boundaries inside it?
Show answer and explanation
Answer: A — Cache checkpointing
Cache checkpointing is specifically about creating reusable cache boundaries within long prompts, so stable sections are reused across repeated requests even when other parts change.
Keep going
More Model Selection & Optimization questions
- A team is 40% over budget on a pipeline with a large, unchanging system prompt and heavy…
- Which task profile best fits routing to the smallest tier?
- What does the context window include?
- A prompt tuned on one model version behaves differently after an upgrade. What is the…
- Before scaling from 1,000 to 500,000 daily requests, what does the blueprint expect first?