CCAO-F Claude Model Fundamentals Practice Question
A developer is building a RAG application and notices Claude 3.5 Sonnet occasionally hallucinates when provided with a large context window. Which architectural adjustment is most effective for improving factual fidelity?
⚠ Common exam trap
Candidates often suggest increasing the model's temperature or adding more context, failing to realize that excessive, irrelevant information actually increases the likelihood of hallucinations in RAG systems.
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
✓
Implement a semantic reranking step to filter out low-relevance chunks before prompt insertion.
Improving factual fidelity in RAG applications relies on reducing the noise-to-signal ratio within the context window. By implementing a reranking step, the developer ensures that only the most relevant document chunks reach the model. Claude performs significantly better when the prompt is constrained to highly pertinent information, as this minimizes the risk of the model prioritizing distractor content or irrelevant retrieval artifacts during the generation phase.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Increase the temperature parameter to 1.0 to allow for more creative exploration.
Why it's wrong here
Higher temperature settings increase the randomness of the model's token selection process. In a RAG context, this creates a higher probability of the model deviating from the provided evidence, thereby exacerbating hallucination risks rather than grounding the output in the retrieved documents.
- ✗
Switch to Claude 3 Haiku to benefit from its faster token processing speeds.
Why it's wrong here
While Haiku offers lower latency, it has a lower capacity for complex reasoning compared to Sonnet or Opus. Switching to a smaller model does not address the underlying issue of context noise and may actually decrease the model's ability to synthesize evidence accurately.
- ✓
Implement a semantic reranking step to filter out low-relevance chunks before prompt insertion.
Why this is correct
Semantic reranking significantly improves RAG performance by ensuring that only the most contextually relevant chunks are fed into the prompt. This reduces the cognitive load on Claude, allowing it to focus on synthesized information rather than filtering through potentially irrelevant or misleading document snippets.
- ✗
Remove the system prompt to allow the model to operate without behavioral constraints.
Why it's wrong here
Removing the system prompt eliminates the primary mechanism for instructing the model on how to handle RAG evidence. A well-defined system prompt is essential for enforcing strict adherence to provided source material and ensuring consistent output formatting throughout the interaction.
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.