Answer at least 13 out of 15 questions correctly to earn your certificate. Questions are randomly selected from a larger bank - each attempt is different.
1.What does Top-P (Nucleus Sampling) control?
2.What is "Self-Consistency" and how does it build on Chain-of-Thought?
3.What is "Function Calling" (tool use) in LLMs?
4.What problem does "Structured Outputs" solve vs. asking for JSON in the prompt?
5.What is "Constraint Stacking"?
6.The course recommends changing only one thing at a time when iterating prompts. Why?
7.When should you escalate from zero-shot to few-shot?
8.What is the "cost-quality ladder" for model selection?
9.What is "Prompt Injection" and how do you defend against it?
10.What temperature should you use for code generation and data extraction?
11.What is an "embedding" in the context of RAG?
12.In a RAG pipeline, what is the correct order of steps?
13.What is "Meta-Prompting"?
14.In the ReAct paradigm, what is the correct loop sequence?
15.Why are positive instructions better than negative ones?