NEW
Font size
WorksheetsGen AI Tech Quiz
Total questions: 25
Worksheet time: 19mins
Which term best describes systems that can learn from data?
Traditional programming
Artificial Intelligence
Machine Learning
Automation
Deep Learning is a subset of:
Data Science
Artificial Intelligence
Machine Learning
Robotics
Which model type uses neural networks with many layers?
Rule-based systems
Deep Learning
Symbolic AI
Expert systems
Which is an example of a multimodal model?
A model that only processes text
A model that only processes images
A model that handles text + images + audio
A model that runs faster
Which stage comes first in the ML lifecycle?
Deployment
Model training
Data collection
Monitoring
Which of the following is an example of Generative AI?
Spam filter
ChatGPT
Calculator
Antivirus
What is the first step in the ML lifecycle?
Model training
Deployment
Problem framing
Testing
What does “training data” mean?
Data used after deployment
Data used to teach the model patterns
Random data
Backup data
Which term refers to fairness and avoiding discrimination in AI?
Speed
Ethics
Accuracy
Automation
What is a “prompt” in Generative AI?
A database query
A user instruction to the model
A training dataset
A deployment file
Which correctly shows the evolution of AI approaches?
Neural → Statistical → Symbolic
Symbolic → Statistical → Neural
Statistical → Neural → Symbolic
Rule-based → Database → Cloud
What is the main difference between ML and DL?
ML uses data, DL does not
DL uses neural networks with many layers
ML is only for images
DL does not require training
Which stage comes after model building in the ML lifecycle?
Data collection
Deployment
Problem framing
Data cleaning
What is “pretraining” in LLMs?
Using the model after deployment
Training the model on large general datasets
Removing biased data
Running prompts
What does RAG (Retrieval-Augmented Generation) do?
Trains a new model
Retrieves external documents before generating answers
Reduces model size
Removes hallucinations completely
Temperature in prompt engineering controls:
Response length
Randomness/creativity of output
Model size
Training speed
A “foundation model” is:
A small dataset
A pre-trained model adaptable to many tasks
A database
A user interface
Which is a real-world use of Generative AI?
Writing emails automatically
Sorting hardware cables
Power management
Network routing
Bias in AI usually comes from:
Internet speed
Training data and design choices
Model deployment
Programming language
Scenario: You want more creative answers from ChatGPT. What should you adjust?
Decrease temperature
Increase temperature
Remove the prompt
Reduce training data
Scenario: A company wants a GenAI system that answers questions using its internal policy documents only. Which technique should be used?
Fine-tuning only
RAG (Retrieval-Augmented Generation)
Data cleaning
Tokenization
Scenario: Your chatbot gives fluent but incorrect answers. This problem is known as:
Overfitting
Bias
Hallucination
Undertraining
Which combination correctly represents the LLM lifecycle?
Data → Deployment → Training → Inference
Pretraining → Fine-tuning → RAG → Inference
Prompting → Testing → Training → Evaluation
Design → Coding → Deployment → Scaling
Scenario: You are building a GenAI system for healthcare. What is the most critical concern?
UI design
Ethics, privacy, and bias
Model speed
Prompt length
Scenario: Two users give the same prompt but receive different outputs. Which parameter most likely caused this?
Token limit
Temperature / Top-P
Dataset size
API key
