Learn. Practice. Feel prepared.
Generative AI Objective Questions and Answers - Set 2
This Generative AI quiz contains carefully curated objective questions with correct answers and clear explanations. It is designed for developers, AI engineers, and interview preparation, covering modern generative models, architectures, optimization techniques, and practical applications. This is part 2 of the series.
Practice Generative AI MCQs with Detailed Explanations
Answer at least 12 questions to submit and verify answers.
16 What is GAN? Medium
Correct answer:
Generative Adversarial Network
GANs use generator and discriminator models.
17 What is discriminator in GAN? Medium
Correct answer:
Evaluates generated data
Discriminator distinguishes real vs fake data.
18 Which improves output diversity? Medium
Correct answer:
High temperature
Higher temperature increases randomness.
19 What is prompt engineering? Medium
Correct answer:
Designing effective prompts
Prompt engineering optimizes model inputs.
20 What is context window? Medium
Correct answer:
Max tokens model can process
Defines input size limit.
21 Which reduces repetition? High
Correct answer:
Frequency penalty
Frequency penalty reduces repeated tokens.
22 What is top-p sampling? High
Correct answer:
Probability threshold sampling
Top-p selects tokens within probability mass.
23 Which is multimodal AI? Medium
Correct answer:
Text + image/audio
Multimodal models process multiple input types.
24 What is overfitting? Medium
Correct answer:
Memorizing training data
Overfitting reduces generalization.
25 Which improves generalization? Medium
Correct answer:
More data
More diverse data improves generalization.
26 What is inference? Medium
Correct answer:
Using model to generate output
Inference is the prediction phase.
27 Which is decoder-only model? High
Correct answer:
GPT
GPT uses decoder-only transformer architecture.
28 What is encoder model? Medium
Correct answer:
Processes input representations
Encoders process input features.
29 Which is encoder-decoder model? High
Correct answer:
T5
T5 uses encoder-decoder architecture.
30 What is RLHF? High
Correct answer:
Reinforcement Learning with Human Feedback
RLHF aligns model outputs with human preferences.