hardMultiple Choice
AIF-C01 Practice Question: An AI practitioner is deploying a large language…
An AI practitioner is deploying a large language model (LLM) for a customer support application. They are concerned about hallucinations, where the model generates plausible but incorrect information. Which combination of techniques would be MOST effective to mitigate hallucinations?
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
✓
Use Retrieval-Augmented Generation (RAG) and enable Bedrock Guardrails
Grounding the model with RAG and using Bedrock Guardrails are effective techniques to reduce hallucinations by providing context and enforcing constraints.
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 Retrieval-Augmented Generation (RAG) and enable Bedrock Guardrails
Why this is correct
RAG grounds responses in retrieved authoritative documents, reducing fabricated content, while Bedrock Guardrails applies content filters and grounding checks that block unsupported claims. Together they address hallucination at both retrieval and output stages, satisfying the mitigation requirement.
- ✗
Disable human review and increase max tokens
Why it's wrong here
Removing human review eliminates the verification layer that catches fabricated output, and raising max tokens simply lengthens responses, giving hallucinations more room to appear. Human review is genuinely useful for high-stakes or low-confidence answers; grounding responses in retrieved source documents is what actually reduces hallucination.
- ✗
Increase model temperature and use top-k sampling
Why it's wrong here
Raising temperature increases randomness and top-k sampling widens the candidate pool, both making fabricated content more likely rather than less. Higher temperature is legitimately used for creative or diverse generation; for factual customer support, lowering temperature and grounding answers in retrieved documents reduces hallucination.
- ✗
Fine-tune the model on a small dataset and reduce context length
Why it's wrong here
Fine-tuning on a small dataset teaches style and format, not factual grounding, and shortening context strips the retrieved evidence the model needs, so hallucinations persist. It tempts because fine-tuning suits narrow tone or task adaptation, and reduced context lowers cost where prompts are already self-contained.
Go deeper
Related to this question
About these practice questions
This AIF-C01 question is part of Courseiva's 862-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 by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This AIF-C01 practice question is part of Courseiva's free Amazon Web Services 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 AIF-C01 exam.