Final Exam: Generative AI Complete
Comprehensive exam covering all 12 modules.
50 questions - Click to reveal answers.
Part 1: Foundations (Questions 1-10)
Q1: What is the primary function of Generative AI?
A) Classify data B) Create new content C) Only analyze D) Delete data
Answer: B - Generative AI creates new content (text, images, audio, video, code).
Q2: What architecture powers most modern LLMs?
A) CNN B) RNN C) Transformer D) Decision tree
Answer: C - Transformers with self-attention are the foundation of modern LLMs.
Q3: What is a token in language models?
A) Payment unit B) Basic text unit processed C) Hardware D) Password
Answer: B - Tokens are the basic units (words, subwords, characters) processed by LLMs.
Q4: What does self-attention allow models to do?
A) Sleep B) Weigh importance of all input parts C) Delete D) Compress
Answer: B - Self-attention lets each part of input consider all other parts.
Q5: What is the transformer's main advantage over RNNs?
A) Smaller B) Parallel processing, handles long sequences C) Older D) Simpler
Answer: B - Transformers can process in parallel and better handle long-range dependencies.
Q6: What is a hallucination in AI?
A) Visual effect B) Generated false information C) Hardware error D) Speed issue
Answer: B - Fabricated information that AI presents as factual.
Q7: What does temperature control?
A) Hardware cooling B) Output randomness/creativity C) Speed D) Size
Answer: B - Temperature controls randomness; low = focused, high = creative.
Q8: What is fine-tuning?
A) Repairing B) Adapting pre-trained model for specific task C) Deleting D) Compressing
Answer: B - Adapting a pre-trained model for specific tasks or domains.
Q9: What year marked the "Generative AI explosion"?
A) 2010 B) 2018 C) 2022 D) 2000
Answer: C - 2022 with ChatGPT's release and widespread public awareness.
Q10: What is a foundation model?
A) Building material B) Large pre-trained model for many tasks C) Basic program D) Hardware
Answer: B - Large model trained on broad data, adaptable for many applications.
Part 2: Text Generation (Questions 11-18)
Q11: What is prompt engineering?
A) Building hardware B) Crafting effective AI instructions C) Programming D) Deleting prompts
Answer: B - Designing prompts to achieve desired outputs from AI models.
Q12: What is Chain-of-Thought prompting?
A) Linking prompts B) Encouraging step-by-step reasoning C) Chain reactions D) Multiple models
Answer: B - Technique that encourages models to show reasoning steps.
Q13: What is RAG?
A) Storage format B) Retrieval-Augmented Generation C) Speed measure D) Error type
Answer: B - Combining language models with external knowledge retrieval.
Q14: What is few-shot learning?
A) Photography B) Learning from few examples in prompt C) Short training D) Quick inference
Answer: B - Learning from a small number of examples provided in the prompt.
Q15: What is context window?
A) UI element B) Maximum text a model can process C) Windows OS D) Visual display
Answer: B - The maximum amount of text a model can consider at once.
Q16: What is a system prompt?
A) System error B) Instructions defining AI behavior C) OS command D) Hardware setting
Answer: B - Hidden instructions that set the AI's behavior and context.
Q17: What is zero-shot prompting?
A) No prompts B) Task completion without examples C) Quick response D) Empty output
Answer: B - Getting AI to perform tasks without providing examples.
Q18: What is tokenization?
A) Payment B) Breaking text into tokens C) Encryption D) Compression
Answer: B - Process of splitting text into the units the model processes.
Part 3: Visual Generation (Questions 19-26)
Q19: How do diffusion models generate images?
A) Direct creation B) Gradually removing noise C) Copying D) Compression
Answer: B - Learning to reverse a noise-adding process, generating from noise to image.
Q20: What is inpainting?
A) Interior design B) Filling in masked image areas C) Painting AI D) Color adjustment
Answer: B - Filling in missing or selected parts of an image.
Q21: What is outpainting?
A) Outdoor painting B) Extending images beyond boundaries C) Removing paint D) Color correction
Answer: B - Generating content to extend an image beyond its original borders.
Q22: What is a negative prompt?
A) Rejection B) Specifying what to avoid in generation C) Negative space D) Error message
Answer: B - Text specifying elements to exclude from generated images.
Q23: What is img2img?
A) Copying B) Transforming input image based on text C) Comparison D) Format
Answer: B - Using an image plus text prompt to guide generation.
Q24: What is CLIP?
A) Hardware B) Model connecting text and images C) Video format D) Shortening tool
Answer: B - Contrastive Language-Image Pre-training, understanding both modalities.
Q25: What is video generation's main challenge?
A) Colors B) Temporal consistency and compute C) Resolution D) Format
Answer: B - Maintaining consistency across frames and intensive computation.
Q26: What is style transfer?
A) Moving files B) Applying artistic style to images C) Font changes D) Layout transfer
Answer: B - Applying the visual style of one image to the content of another.
Part 4: Audio and Code (Questions 27-34)
Q27: What is voice cloning?
A) Duplicating hardware B) Creating synthetic speech from voice samples C) Recording D) Editing
Answer: B - Generating speech that mimics a specific person's voice.
Q28: What is GitHub Copilot?
A) Version control B) AI coding assistant C) Repository host D) Bug tracker
Answer: B - AI pair programmer that suggests code completions and functions.
Q29: What does AI code review do?
A) Delete code B) Analyze code for issues and improvements C) Execute code D) Compile code
Answer: B - Identifies bugs, security issues, and suggests improvements.
Q30: What is a limitation of AI code generation?
A) Too fast B) May produce bugs or insecure code C) Too accurate D) Too small
Answer: B - Can generate non-functional or insecure code requiring human review.
Q31: What is text-to-speech?
A) Transcription B) Converting text to spoken audio C) Translation D) Dictation
Answer: B - Generating natural-sounding speech from written text.
Q32: What is a code agent?
A) Programmer B) AI that writes, tests, and debugs code autonomously C) IDE D) Compiler
Answer: B - AI system that can autonomously perform multi-step coding tasks.
Q33: What is AI music generation?
A) Streaming B) Creating original music with AI C) Mixing D) Recording
Answer: B - Using AI to compose melodies, harmonies, and full songs.
Q34: What is speech-to-text?
A) TTS B) Converting spoken audio to text C) Recording D) Translation
Answer: B - Transcribing spoken language into written text.
Part 5: Multimodal and Business (Questions 35-42)
Q35: What is multimodal AI?
A) Multiple modes B) AI processing multiple data types together C) Multi-user D) Multiple models
Answer: B - AI systems that understand and generate across text, image, audio, etc.
Q36: What is cross-modal generation?
A) Crossing boundaries B) Creating one modality from another C) Switching modes D) Comparison
Answer: B - Generating content in one modality from input in another (text→image).
Q37: What is a key business benefit of Generative AI?
A) More employees B) Automation and efficiency gains C) Higher costs D) Slower processes
Answer: B - Automating tasks, increasing productivity, and enabling new capabilities.
Q38: What is enterprise AI integration?
A) Separate systems B) Connecting AI to business workflows C) Removing AI D) Manual processes
Answer: B - Incorporating AI capabilities into existing business systems and processes.
Q39: What is AI-powered customer service?
A) Phone support B) AI handling customer inquiries C) Email only D) Self-service only
Answer: B - Using AI to handle, triage, or assist with customer interactions.
Q40: What ROI consideration matters for AI projects?
A) Only upfront cost B) Total value vs. total cost over time C) Only speed D) Only features
Answer: B - Evaluating complete costs against quantified benefits over time.
Q41: What is AI content at scale?
A) Large files B) Generating large volumes of content efficiently C) Big screens D) Heavy computing
Answer: B - Using AI to create many pieces of content for marketing, etc.
Q42: What is prompt injection?
A) Medical procedure B) Manipulating AI through malicious prompts C) Code injection D) Data entry
Answer: B - Security attack using crafted prompts to manipulate AI behavior.
Part 6: Ethics and Future (Questions 43-50)
Q43: What is AI bias?
A) Preference B) Systematic unfairness in AI outputs C) Opinion D) Speed
Answer: B - Systematic errors causing unfair outcomes for certain groups.
Q44: What is deepfake?
A) Deep learning B) AI-generated realistic fake media C) Deep analysis D) Deep storage
Answer: B - Synthetic media where a person's likeness is artificially manipulated.
Q45: What is AI governance?
A) Government AI B) Policies and oversight for AI use C) AI control D) AI voting
Answer: B - Frameworks, policies, and processes for responsible AI management.
Q46: What is the EU AI Act?
A) Performance B) Comprehensive AI regulation framework C) Hardware standard D) Protocol
Answer: B - First comprehensive legal framework for AI regulation.
Q47: What is AI transparency?
A) Invisible AI B) Clear communication about AI use and limitations C) See-through hardware D) Open source
Answer: B - Being open about when AI is used and how it works.
Q48: What is AGI?
A) Company B) Human-level AI across all domains C) Model type D) Framework
Answer: B - Artificial General Intelligence matching human intelligence broadly.
Q49: What is the alignment problem?
A) Text formatting B) Ensuring AI acts according to human values C) Model training D) Data alignment
Answer: B - Challenge of ensuring AI systems behave as humans intend.
Q50: What skill is essential for the AI future?
A) Avoiding AI B) AI literacy and continuous learning C) Only coding D) Only theory
Answer: B - Understanding AI and continuously adapting as the field evolves.
Scoring
| Score | Grade | Level |
|---|---|---|
| 45-50 | A | Excellent - Ready to apply Generative AI professionally |
| 40-44 | B | Good - Strong foundation, some areas to strengthen |
| 35-39 | C | Satisfactory - Review weaker modules |
| 30-34 | D | Needs improvement - More study recommended |
| Below 30 | F | Insufficient - Comprehensive review needed |
Certificate
Scoring 45+ on this exam demonstrates comprehensive understanding of Generative AI concepts and applications.
What's Next?
- Review the Glossary for term refreshers
- Explore Additional Resources for continued learning
- Apply your knowledge in real projects!
Congratulations on completing the Generative AI Complete Course!