Generative AI

← All exam tracks

LLMs · RAG · Responsible AI

Generative AI

Practice original scenario-based Generative AI questions. Take a timed mock, focus on a topic, and learn from detailed explanations. These are independent practice materials, not official exam questions or a guarantee of certification readiness.

16 questions · 25 minutes · 70% practice target

Create an account to practice Already a member? Sign in

Topics you will practice

  • Foundations
  • Prompting & RAG
  • Building & Evaluation
  • Safety & Operations

Try a sample

Think through each question, then reveal the explanation. Full attempts save your answers and scores to your account.

A long document exceeds a model context window. What is the most direct consequence?

  1. The request must be truncated, summarized, or split
  2. The model automatically stores it permanently
  3. Temperature becomes zero
  4. Embeddings become unnecessary
Show answer and explanation

Correct answer: The request must be truncated, summarized, or split

The context window limits input plus output tokens; oversized material needs selection, chunking, or summarization.

A team needs more repeatable classification output from the same prompt. Which change helps most?

  1. Raise temperature
  2. Lower temperature and constrain the output schema
  3. Add unrelated examples
  4. Remove the task instruction
Show answer and explanation

Correct answer: Lower temperature and constrain the output schema

Lower randomness and a constrained schema improve repeatability for structured tasks.

What do embeddings primarily represent?

  1. Encrypted source text
  2. Semantic features as numeric vectors
  3. A model password
  4. Guaranteed factual truth
Show answer and explanation

Correct answer: Semantic features as numeric vectors

Embeddings map content into vectors where distance can approximate semantic similarity.