Cyberlium

AI › Module 4 › Lesson 5

BeginnerModule 4Lesson 5/5

Quiz — Poisoning

10 APPLY items on data poisoning, adversarial ML, and supply chain on $AI_LAB.

10 min+40 XP10 quiz
Module progress5 of 5

Opening

Poisoning and Adversarial — Module Quiz

Ten APPLY items. Name poisoning and adversarial concepts on YOUR $AI_LAB — never stranger pipeline sabotage, prod API fuzzing, or unverified weight imports. Original Cyberlium. Next: AI Phishing Named.

Knowledge Check

1

APPLY: Data poisoning corrupts:

Multiple choice

Knowledge Check

2

APPLY: True or False: Backdoor triggers can activate only under attacker-chosen conditions.

True or False

Knowledge Check

3

APPLY: Evasion attacks operate primarily at:

Multiple choice

Knowledge Check

4

APPLY: Tutorial uploads poison to public model hub without scope — you:

Multiple choice

Knowledge Check

5

APPLY: Model supply chain checklist includes:

Multiple choice

Knowledge Check

6

APPLY: Cyberlium NEVER allows:

Multiple choice

Knowledge Check

7

APPLY: Adversarial robustness testing should use:

Multiple choice

Knowledge Check

8

APPLY: True or False: RAG corpus can be a poisoning target separate from base model weights.

True or False

Knowledge Check

9

APPLY: chmod 600 on poisoning notes means:

Multiple choice

Knowledge Check

10

APPLY: Next lesson after this quiz:

Multiple choice

← Previous

Answer all 10 knowledge checks to continue. (0/10 answered)