100% Latest & Updated Google Generative AI Leader Practice Test Questions, Exam Dumps & Verified Answers!
30 Days Free Updates, Instant Download!
Question 610
A new prompt seems promising in one demo but fails on unusual inputs. What should precede rollout?
- Examples and counterexamples
- Explicit output format
- Iterative prompt evaluation
- Clear task objective
Explanation
Correct Answer
C:Representative evaluation exposes brittle instructions. Iteration tests alternative instructions against representative inputs.
Incorrect Answers
A:Examples illustrate acceptable and unacceptable outputs when wording alone is ambiguous. This does not directly meet the case requirement to test prompts on diverse relevant cases.
B:Format instructions specify length, headings, schema or layout expectations. This does not directly meet the case requirement to test prompts on diverse relevant cases.
D:A clear objective tells the model the exact requested action and intended outcome. This does not directly meet the case requirement to test prompts on diverse relevant cases.
Question 611
A consultant asks a model to explain a complex topic for nontechnical executives. What prompt technique can set the audience?
- External evidence verification
- Role prompting
- Stepwise task decomposition
- Structured final answer
Explanation
Correct Answer
B:Role instructions can shape expertise and communication level. Role prompts specify the response persona and relevant expertise or audience.
Incorrect Answers
A:Verification checks factual claims against accessible sources rather than assuming internal reasoning is correct. This does not directly meet the case requirement to adopt a professional executive-facing voice.
C:Decomposition organizes a hard problem into smaller explicit subproblems. This does not directly meet the case requirement to adopt a professional executive-facing voice.
D:A structured final answer presents results in a user-facing format without revealing internal hidden reasoning. This does not directly meet the case requirement to adopt a professional executive-facing voice.
Question 612
An AI process drafts a list of risks, then separately ranks them and creates a memo. What pattern combines the stages?
- Stepwise task decomposition
- External evidence verification
- Structured final answer
- Prompt chaining
Explanation
Correct Answer
D:Chaining passes intermediate products to later prompts. Prompt chaining uses multiple successive tasks whose outputs feed later stages.
Incorrect Answers
A:Decomposition organizes a hard problem into smaller explicit subproblems. This does not directly meet the case requirement to run multiple prompts in an ordered workflow.
B:Verification checks factual claims against accessible sources rather than assuming internal reasoning is correct. This does not directly meet the case requirement to run multiple prompts in an ordered workflow.
C:A structured final answer presents results in a user-facing format without revealing internal hidden reasoning. This does not directly meet the case requirement to run multiple prompts in an ordered workflow.
Question 613
A support agent alternates deciding what information is missing, calling a search tool and using the result to choose a next action. Which pattern is closest?
- ReAct-style tool loop
- Role prompting
- External evidence verification
- Stepwise task decomposition
Explanation
Correct Answer
A:ReAct combines tool actions and observations in an iterative workflow. ReAct patterns coordinate reasoning-like planning and external actions or observations through tools.
Incorrect Answers
B:Role prompts specify the response persona and relevant expertise or audience. This does not directly meet the case requirement to coordinate decisions with external tool observations.
C:Verification checks factual claims against accessible sources rather than assuming internal reasoning is correct. This does not directly meet the case requirement to coordinate decisions with external tool observations.
D:Decomposition organizes a hard problem into smaller explicit subproblems. This does not directly meet the case requirement to coordinate decisions with external tool observations.
Question 614
A complex data analysis is separated into cleaning, calculations and business interpretation. What planning technique helps?
- External evidence verification
- Stepwise task decomposition
- ReAct-style tool loop
- Prompt chaining
Explanation
Correct Answer
B:Task decomposition reduces the number of concerns per step. Decomposition organizes a hard problem into smaller explicit subproblems.
Incorrect Answers
A:Verification checks factual claims against accessible sources rather than assuming internal reasoning is correct. This does not directly meet the case requirement to break a difficult task into smaller actions.
C:ReAct patterns coordinate reasoning-like planning and external actions or observations through tools. This does not directly meet the case requirement to break a difficult task into smaller actions.
D:Prompt chaining uses multiple successive tasks whose outputs feed later stages. This does not directly meet the case requirement to break a difficult task into smaller actions.
Question 615
A client wants only the result, assumptions and cited sources, not the assistant's private reasoning trace. What output design fits?
- Structured final answer
- Prompt chaining
- ReAct-style tool loop
- External evidence verification
Explanation
Correct Answer
A:Structured answers can provide verifiable conclusions succinctly. A structured final answer presents results in a user-facing format without revealing internal hidden reasoning.
Incorrect Answers
B:Prompt chaining uses multiple successive tasks whose outputs feed later stages. This does not directly meet the case requirement to present results without hidden chain-of-thought disclosure.
C:ReAct patterns coordinate reasoning-like planning and external actions or observations through tools. This does not directly meet the case requirement to present results without hidden chain-of-thought disclosure.
D:Verification checks factual claims against accessible sources rather than assuming internal reasoning is correct. This does not directly meet the case requirement to present results without hidden chain-of-thought disclosure.
Question 616
A fluent assistant asserts a legal citation based on its own internal steps but has not checked the document. Which activity is still needed?
- External evidence verification
- Role prompting
- Stepwise task decomposition
- Prompt chaining
Explanation
Correct Answer
A:Reasoning quality does not replace factual source verification. Verification checks factual claims against accessible sources rather than assuming internal reasoning is correct.
Incorrect Answers
B:Role prompts specify the response persona and relevant expertise or audience. This does not directly meet the case requirement to verify evidence for consequential claims.
C:Decomposition organizes a hard problem into smaller explicit subproblems. This does not directly meet the case requirement to verify evidence for consequential claims.
D:Prompt chaining uses multiple successive tasks whose outputs feed later stages. This does not directly meet the case requirement to verify evidence for consequential claims.
Question 617
A model is instructed to answer as a calm customer support representative. What is the named prompt pattern?
- Prompt chaining
- Role prompting
- External evidence verification
- ReAct-style tool loop
Explanation
Correct Answer
B:Role prompting specifies the response posture. Role prompts specify the response persona and relevant expertise or audience.
Incorrect Answers
A:Prompt chaining uses multiple successive tasks whose outputs feed later stages. This does not directly meet the case requirement to tailor voice to a support role.
C:Verification checks factual claims against accessible sources rather than assuming internal reasoning is correct. This does not directly meet the case requirement to tailor voice to a support role.
D:ReAct patterns coordinate reasoning-like planning and external actions or observations through tools. This does not directly meet the case requirement to tailor voice to a support role.
Question 618
A content workflow separately summarizes evidence, checks dates and writes a final brief. Which prompting technique is involved?
- ReAct-style tool loop
- Stepwise task decomposition
- External evidence verification
- Prompt chaining
Explanation
Correct Answer
D:Chained steps help manage multi-stage work. Prompt chaining uses multiple successive tasks whose outputs feed later stages.
Incorrect Answers
A:ReAct patterns coordinate reasoning-like planning and external actions or observations through tools. This does not directly meet the case requirement to feed intermediate summaries into subsequent prompts.
B:Decomposition organizes a hard problem into smaller explicit subproblems. This does not directly meet the case requirement to feed intermediate summaries into subsequent prompts.
C:Verification checks factual claims against accessible sources rather than assuming internal reasoning is correct. This does not directly meet the case requirement to feed intermediate summaries into subsequent prompts.
Question 619
An agent needs to decide between searching an index and calling a pricing API as new evidence arrives. What style describes the loop?
- Role prompting
- ReAct-style tool loop
- Prompt chaining
- External evidence verification
Explanation
Correct Answer
B:ReAct-style action/observation feedback helps choose tools. ReAct patterns coordinate reasoning-like planning and external actions or observations through tools.
Incorrect Answers
A:Role prompts specify the response persona and relevant expertise or audience. This does not directly meet the case requirement to alternate tool actions and observations.
C:Prompt chaining uses multiple successive tasks whose outputs feed later stages. This does not directly meet the case requirement to alternate tool actions and observations.
D:Verification checks factual claims against accessible sources rather than assuming internal reasoning is correct. This does not directly meet the case requirement to alternate tool actions and observations.
Question 620
A reasoning problem is too large for one response and is divided into a short list of independently checkable questions. What is this called?
- Stepwise task decomposition
- Structured final answer
- External evidence verification
- Prompt chaining
Explanation
Correct Answer
A:Decomposition helps make complex work manageable. Decomposition organizes a hard problem into smaller explicit subproblems.
Incorrect Answers
B:A structured final answer presents results in a user-facing format without revealing internal hidden reasoning. This does not directly meet the case requirement to solve a complex task as component problems.
C:Verification checks factual claims against accessible sources rather than assuming internal reasoning is correct. This does not directly meet the case requirement to solve a complex task as component problems.
D:Prompt chaining uses multiple successive tasks whose outputs feed later stages. This does not directly meet the case requirement to solve a complex task as component problems.
Generative AI Leader Practice Tests
