Learn. Practice. Feel prepared.
GenAI & LLM Engineering Objective Questions and Answers - Set 3
This GenAI & LLM engineering quiz contains carefully curated objective questions with correct answers and clear explanations. It is designed for developers and ML engineers to test your skills in building, optimizing, evaluating, and deploying LLM-powered applications. This is part 3 of the series.
Practice GenAI & LLM Engineering MCQs with Detailed Explanations
Answer at least 12 questions to submit and verify answers.
31 Which scenario benefits most from function calling or tool calling? Medium
Correct answer:
Structured actions like booking or calculations
Tool calling enables structured outputs for deterministic actions like APIs or calculations.
32 What is the primary challenge with long-term memory in LLM agents? Hard
Correct answer:
Maintaining relevance and avoiding stale context
Long-term memory must stay relevant and avoid polluting context with outdated or irrelevant data.
33 Which approach best supports multi-turn conversation consistency? Medium
Correct answer:
Conversation state management with summarization
Summarizing and managing conversation state preserves context across turns efficiently.
34 What is a major drawback of beam search for open-ended generation? Hard
Correct answer:
Reduced diversity and repetitive outputs
Beam search often reduces diversity and can produce repetitive or generic outputs.
35 Which factor most influences embedding quality for domain-specific retrieval? Medium
Correct answer:
Embedding model choice and domain adaptation
Choosing or adapting embedding models to the domain significantly impacts retrieval quality.
36 What is the main risk of storing raw user prompts for training? Hard
Correct answer:
Privacy and compliance issues
Storing raw prompts can expose sensitive user data and violate privacy or compliance requirements.
37 Which technique helps reduce token usage in long conversations? Medium
Correct answer:
Conversation summarization
Summarization condenses prior context, reducing token usage while preserving key information.
38 What is the main benefit of using evaluation harnesses for LLMs? Hard
Correct answer:
Systematic and repeatable benchmarking
Evaluation harnesses provide repeatable, automated testing across tasks and prompts.
39 Which method best reduces cost for high-volume LLM usage? Medium
Correct answer:
Caching + smaller distilled models
Caching frequent responses and using smaller distilled models reduces inference cost.
40 What is the key purpose of guardrails in LLM applications? Medium
Correct answer:
Enforce safety, policy, and output constraints
Guardrails enforce safety, policy compliance, and structured output constraints.
41 Which factor most affects vector search accuracy at scale? Hard
Correct answer:
Index type and distance metric
Index structures and distance metrics directly impact retrieval accuracy and performance.
42 What is the main trade-off of aggressive vector compression? Hard
Correct answer:
Lower memory but reduced recall
Compression reduces memory usage but can degrade retrieval recall and accuracy.
43 Which approach improves reliability of LLM outputs in critical workflows? Hard
Correct answer:
Multi-step reasoning with verification
Verification and multi-step reasoning improve reliability for critical tasks.
44 What is the main risk of tool hallucination in agents? Hard
Correct answer:
Calling non-existent tools or wrong parameters
Agents may attempt to call tools incorrectly or invent tools without proper validation.
45 Which technique best supports multilingual retrieval in RAG? Medium
Correct answer:
Multilingual embedding models
Multilingual embeddings align semantic meaning across languages for retrieval.