CCAR-F Prompt Engineering and Structured Output Practice Question
You are iterating on a prompt that summarizes technical articles. The summaries are sometimes too long and sometimes miss the required 'key_takeaways' section. You want to diagnose whether the problem is the instruction wording or the input content. Which practice best supports systematic prompt iteration?
⚠ Common exam trap
The trap here is believing that a stronger model or self-critique replaces a repeatable evaluation set, when only controlled comparison across versions reveals the true cause.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Maintain a small set of representative articles with expected outputs and run the prompt against them after each change.
Using a fixed evaluation set with expected outputs is the core of systematic prompt iteration. It makes results comparable across versions and isolates the effect of instruction wording from input content. Changing many things at once, switching models, or trusting self-critique does not provide the controlled measurement needed to diagnose the length and missing-section problems.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✓
Maintain a small set of representative articles with expected outputs and run the prompt against them after each change.
Why this is correct
A fixed evaluation set with expected outputs lets you compare results across prompt versions and see whether a change improves length control or key_takeaways coverage. In this scenario, it separates wording effects from content effects because the same inputs are reused. This is the foundation of reliable prompt iteration and regression testing.
- ✗
Change several instructions at once and evaluate the final output by reading it once.
Why it's wrong here
Changing multiple variables simultaneously makes it impossible to attribute an improvement or regression to a specific change. Reading a single output is anecdotal and does not reveal whether the issue is wording or content. In this scenario, that approach would obscure the diagnosis rather than support it. Systematic iteration requires controlled, isolated changes.
- ✗
Switch to a larger model and assume the summaries will meet the requirements.
Why it's wrong here
Model size does not fix an ambiguous prompt or a missing section requirement. A larger model may follow instructions better, but without measurement you cannot know if the issue is wording or content. In this scenario, this change adds cost without diagnostic value. It does not support systematic iteration.
- ✗
Ask Claude to critique its own summary and trust the critique as the final answer.
Why it's wrong here
Self-critique can be useful, but it is not a substitute for a stable evaluation set. The model may miss the same issues it produced, and the critique is not comparable across prompt versions. In this scenario, relying on it alone would not reveal whether wording or input content is the cause. It also consumes extra tokens without a clear baseline.
About these practice questions
This CCAR-F question is part of Courseiva's 271-question bank — original exam-style content with full explanations and wrong-answer analysis, never real exam questions or exam dumps. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official Anthropic exam blueprint
This CCAR-F practice question is part of Courseiva's free Anthropic certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the CCAR-F exam.