CCAO-F Claude Model Fundamentals Practice Question
An engineering team wants Claude to classify thousands of support tickets into categories. They need the model to always return one of five exact category labels. Which approach most reliably constrains the output to those labels?
⚠ Common exam trap
The trap here is relying on post-hoc parsing or vague wording to fix classification output, when the robust solution is to define the exact allowed labels and constrain generation up front.
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
✓
List the five allowed labels in the system prompt and instruct the model to output only one of them.
To force output into a closed set of labels, the prompt must enumerate the allowed values and instruct the model to return exactly one. Placing this in the system prompt ensures the constraint applies to every classification request. Vague instructions, high temperature, or downstream keyword matching all allow invalid or ambiguous outputs that break automated pipelines.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Use a high temperature so the model explores all categories evenly.
Why it's wrong here
High temperature increases randomness and can cause the model to produce unexpected tokens, including labels outside the allowed set. Classification into a fixed taxonomy benefits from low temperature, where the most probable label is chosen consistently. Exploration is counterproductive when the goal is strict adherence to five predefined categories.
- ✓
List the five allowed labels in the system prompt and instruct the model to output only one of them.
Why this is correct
Enumerating the exact allowed labels and instructing the model to return only one of them gives a precise constraint. The system prompt applies to every request, so the model consistently sees the closed set. This directly satisfies the requirement that outputs match one of five fixed strings, making downstream parsing reliable.
- ✗
Post-process the model's free-text output with a keyword search for category names.
Why it's wrong here
Post-processing free text is brittle: the model may mention multiple category names in its explanation, and a keyword search could match the wrong one or none at all. This adds complexity and failure modes instead of preventing invalid outputs at the source. Constraining generation directly is more reliable than trying to repair unconstrained text afterward.
- ✗
Ask the model to 'pick the best category' in the user message.
Why it's wrong here
A vague instruction like 'pick the best category' does not define the allowed set or enforce exact labels. Claude might invent a category, return a synonym, or include extra commentary. Without an explicit enumeration and formatting rule, outputs will vary, breaking downstream automation that expects one of five fixed strings.
About these practice questions
One of 259 original CCAO-F practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. 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 CCAO-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 CCAO-F exam.