CCAR-F Prompt Engineering and Structured Output Practice Question
Which THREE strategies are effective for reducing hallucinations in RAG (Retrieval-Augmented Generation) systems?
⚠ Common exam trap
Test-takers often assume that simply retrieving documents via RAG is enough, forgetting that models still tend to extrapolate beyond provided context.
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
✓
Instruct the model to say 'I don't know' if the context does not contain the answer.
To reduce hallucinations in RAG, you must constrain the model to the provided context, encourage it to admit ignorance if the answer is missing, and verify its reasoning. These techniques transform the model from a creative writer into a grounded processor of provided facts. Architects must build these guardrails into the prompt to ensure the output is factually tethered to the retrieved source material, which is critical for trustworthy systems.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✓
Instruct the model to say 'I don't know' if the context does not contain the answer.
Why this is correct
Explicitly instructing the model to admit ignorance prevents it from filling gaps with hallucinated information. This is a critical guardrail for RAG systems, ensuring that accuracy is prioritized over completing the task, which helps maintain user trust and factual integrity in enterprise-grade information retrieval applications.
- ✓
Force the model to cite the specific document chunk used for each sentence.
Why this is correct
Requiring citations forces the model to ground its response in the provided context. This makes verification easier for the user and discourages the model from deviating from the source material. It is a standard practice for improving the accuracy and auditability of RAG pipelines in production environments.
- ✗
Increase the temperature to 1.5 to allow for more creative exploration.
Why it's wrong here
High temperature settings increase the probability of hallucinations, which is detrimental to RAG systems. RAG requires high factual adherence, which is best achieved by lower temperatures. Increasing the temperature will only undermine the accuracy of the system by encouraging the model to generate non-factual, creative content.
- ✗
Use a complex system prompt that describes the model as an 'Omniscient Expert'.
Why it's wrong here
Calling a model 'Omniscient' is counter-productive because it encourages the model to generate information even when it lacks the necessary data in the context. This persona promotes overconfidence and increases hallucination risks. Instead, define the model as a helpful assistant that strictly uses the provided context.
- ✓
Ask the model to verify its own answer against the context provided.
Why this is correct
Self-verification is a powerful technique where the model compares its drafted response to the source material before finalizing. This step allows the model to identify inconsistencies or hallucinations, leading to more grounded and accurate responses that align perfectly with the factual content retrieved from the underlying knowledge base.
About these practice questions
Courseiva writes every CCAR-F question from scratch — 271 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or 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.