Font size
WorksheetsPre-presentation quiz
Total questions: 19
Worksheet time: 10mins
What does "perturbation" mean in the context of adversarial examples?
A large, random modification that changes the meaning of the input
A small, intentional change made to an input to mislead a model
A type of data used to improve model training
A method to visualize model performance
What is the main purpose of applying a perturbation to an input?
To make the input more realistic
To improve the accuracy of the model
To make the model produce an incorrect output
To reduce overfitting
What are vulnerabilities in AI systems?
Simple programming errors or syntax mistakes
Structural weaknesses that cause incorrect or exploitable outputs
Network latency problems in AI hardware
Overfitting issues during training
Why do overfitting, memorization, and brittleness make AI “vulnerable”?
Because people don’t use Wi-Fi
Because AI wants to rest
Because the AI doesn’t actually understand — it just copies patterns
Because AI hates learning
Do AI vulnerabilities exist at the intersection of model architecture, training data quality, and deployment context?
Yes
No
These are characteristics of LLM hallucinations except:
Generating fluent and coherent text
Producing factually incorrect or fabricated content
Always caused by software or coding bugs
Are deepfakes and hallucinations different because of human involvement?
Yes
No
Deepfakes does not involve human manipulation
Yes
No
Models purposely lie when they hallucinate, generating factually incorrect outputs through probabilistic mechanisms without intentional deception
Yes
No
What is intrinsic hallucination?
Output that is intentionally misleading by human manipulation
Output that contains spelling or grammar errors
Output that is fluent and coherent
Output that contradicts the source content
Faithfulness refers to when models stay consistent and truthful to the provided sources
Yes
No
What is the acceptable tolerance level of hallucination in abstractive summarization?
High
Moderate
Low
Very low
Hallucination in open-domain dialogue generation are generally unacceptable, unless they are minor and do not involve sever factual issues
Yes
No
Hallucination in Data-to-Text occurs due to the gap between structured data and the natural language text produced during the conversion
Yes
No
What factors contribute to LLM hallucination from data?
Source-Reference Divergence
Heuristic Data Collection
Dataset Duplication
Innate Divergence
Hallucinations persist partly because current evaluation encourages guessing rather than honesty about uncertainty
Yes
No
If an AI memorizes private information from its data, what could go wrong?
It might share secrets by accident
It becomes shy
It learns faster
It refuses to answer
What are some real-world consequences of misinformation spreading online?
Public confusion
Political polarization
Reputational harm
Moral panic
Why is AI-generated misinformation particularly dangerous compared to traditional misinformation?
It spreads slower and is easier to detect
It often looks and sounds more credible
It can only be created by experts with coding skills
It usually contains obvious factual errors
