WGU D685 EXAM
WGU D685: Practical Applications of Prompt
Engineering — Comprehensive Objective Assessment
With 200 Exam-Style Questions, Detailed Answer
Keys, and In-Depth Rationales Covering Foundations,
Advanced Prompt Crafting, AI Capabilities and
Limitations, Ethical AI and Bias Mitigation, Legal and
Regulatory Compliance, and Evaluation, Iteration,
and Performance Optimization GRADED A+
================================================================================
WGU D685: PRACTICAL APPLICATIONS OF PROMPT ENGINEERING
COMPREHENSIVE OBJECTIVE ASSESSMENT — 200 QUESTIONS
WITH COMPLETE ANSWER KEYS AND RATIONALES
================================================================================
COURSE CODE ............ WGU D685
EXAM TITLE ............. Comprehensive Objective Assessment
VERSION ................ 2026/2027 Edition
TOTAL ITEMS ............ 200
QUESTION TYPES ......... Multiple-Choice, Select-All-That-Apply,
True/False, Case Study, Scenario-Based
RECOMMENDED TIME ....... 3–4 Hours
RECOMMENDED PASS MARK .. 80% (160/200)
================================================================================
SECTION MAP
================================================================================
SECTION I ............ Foundations of Prompt Engineering (Q1–Q35)
SECTION II ............ Advanced Prompt Crafting & Optimization (Q36–Q70)
SECTION III ............ AI Capabilities, Limitations & Hallucination
pg. 1
, WGU D685 EXAM
Management (Q71–Q105)
SECTION IV ............ Ethical AI, Bias Detection & Responsible Prompt
Design (Q106–Q140)
SECTION V ............ Legal & Regulatory Compliance (Q141–Q170)
SECTION VI ............ Evaluation, Iteration & Performance
Optimization (Q171–Q200)
================================================================================
COLOR-CODED FORMAT KEY
================================================================================
[BLUE] QUESTION heading and question section
[RED] ANSWER heading and answer section
[PURPLE] RATIONALE heading and rationale section
================================================================================
EXAM INSTRUCTIONS
================================================================================
1. Read each question carefully before selecting your answer.
2. For multiple-choice items, choose the single best answer.
3. For select-all-that-apply items, choose ALL correct options.
4. Review the Answer and Rationale for every question.
5. Do not skip any question — all 200 items include explanations.
6. Correct answer positions are randomized throughout the exam.
================================================================================
BEGIN EXAM
================================================================================
pg. 2
, WGU D685 EXAM
WGU D685: Practical Applications of Prompt Engineering
Comprehensive Objective Assessment – 200 Questions
Exam Instructions: This examination contains 200 questions covering foundational prompt
engineering principles, advanced prompt crafting techniques, AI model capabilities and limitations,
ethical AI and bias mitigation, legal and regulatory compliance, evaluation and iteration of prompt
performance, and scenario-based problem solving. Select the best answer for each question. Each
question includes a complete answer and rationale.
SECTION I: FOUNDATIONS OF PROMPT ENGINEERING
Questions 1–35
QUESTION 1:
What is the primary goal of prompt engineering in generative AI?
A. To write longer prompts with more details
B. To maximize the likelihood of a desired model output
C. To reduce the number of model parameters
D. To train the model from scratch using custom datasets
ANSWER:
B. To maximize the likelihood of a desired model output
RATIONALE:
Prompt engineering focuses on designing inputs that guide a generative AI model to produce
accurate, relevant, and useful outputs without modifying the model's weights or architecture. The
goal is to elicit the best possible response through careful input design, not to alter the underlying
model. Option A is incorrect because longer prompts are not inherently better. Option C is incorrect
because prompt engineering does not modify model parameters. Option D is incorrect because
training models from scratch is a separate process entirely.
QUESTION 2:
Which of the following is an example of a zero-shot prompt?
A. "Translate 'Hello' to Spanish." (without providing examples)
B. "Here is an example: 'Cat → Gato'. Now translate 'Dog'."
C. "Classify this sentiment as positive or negative after reading 5 examples."
D. "Generate a story based on the following three examples of fairy tales."
ANSWER:
A. "Translate 'Hello' to Spanish." (without providing examples)
RATIONALE:
Zero-shot prompting means the model receives no examples of the task and must rely solely on its
pg. 3
, WGU D685 EXAM
pre-trained knowledge to generate a response. Options B, C, and D all provide examples to the
model, which makes them few-shot prompts rather than zero-shot prompts.
QUESTION 3:
What is the core problem with the prompt: "Where should I eat dinner?"
A. The AI lacks sufficient training data on restaurants
B. The prompt lacks location, cuisine preference, and budget constraints
C. The AI does not understand the concept of dinner
D. The prompt is too long and complex
ANSWER:
B. The prompt lacks location, cuisine preference, and budget constraints
RATIONALE:
This prompt is too vague and lacks essential context that would help the AI provide a useful
response. Without location, cuisine preferences, budget, or other constraints, the AI cannot provide
a relevant recommendation. The AI has sufficient training data on restaurants, understands the
concept of dinner, and the prompt is actually too short and underspecified rather than too long.
QUESTION 4:
Which technique best improves multi-step reasoning in large language models?
A. Role prompting
B. Chain-of-thought (CoT) prompting
C. Length penalty adjustment
D. Temperature adjustment
ANSWER:
B. Chain-of-thought (CoT) prompting
RATIONALE:
Chain-of-thought prompting encourages the model to show intermediate reasoning steps, improving
accuracy for math, logic, and complex reasoning tasks by 40-100% according to research. Role
prompting assigns a persona but does not specifically improve multi-step reasoning. Length penalty
and temperature adjustments affect output length and randomness respectively, but do not directly
enhance reasoning capabilities.
QUESTION 5:
What does the "temperature" parameter control in LLM generation?
A. Length of output
B. Randomness or creativity of outputs
C. Grammar correctness
D. Training data size
pg. 4
WGU D685: Practical Applications of Prompt
Engineering — Comprehensive Objective Assessment
With 200 Exam-Style Questions, Detailed Answer
Keys, and In-Depth Rationales Covering Foundations,
Advanced Prompt Crafting, AI Capabilities and
Limitations, Ethical AI and Bias Mitigation, Legal and
Regulatory Compliance, and Evaluation, Iteration,
and Performance Optimization GRADED A+
================================================================================
WGU D685: PRACTICAL APPLICATIONS OF PROMPT ENGINEERING
COMPREHENSIVE OBJECTIVE ASSESSMENT — 200 QUESTIONS
WITH COMPLETE ANSWER KEYS AND RATIONALES
================================================================================
COURSE CODE ............ WGU D685
EXAM TITLE ............. Comprehensive Objective Assessment
VERSION ................ 2026/2027 Edition
TOTAL ITEMS ............ 200
QUESTION TYPES ......... Multiple-Choice, Select-All-That-Apply,
True/False, Case Study, Scenario-Based
RECOMMENDED TIME ....... 3–4 Hours
RECOMMENDED PASS MARK .. 80% (160/200)
================================================================================
SECTION MAP
================================================================================
SECTION I ............ Foundations of Prompt Engineering (Q1–Q35)
SECTION II ............ Advanced Prompt Crafting & Optimization (Q36–Q70)
SECTION III ............ AI Capabilities, Limitations & Hallucination
pg. 1
, WGU D685 EXAM
Management (Q71–Q105)
SECTION IV ............ Ethical AI, Bias Detection & Responsible Prompt
Design (Q106–Q140)
SECTION V ............ Legal & Regulatory Compliance (Q141–Q170)
SECTION VI ............ Evaluation, Iteration & Performance
Optimization (Q171–Q200)
================================================================================
COLOR-CODED FORMAT KEY
================================================================================
[BLUE] QUESTION heading and question section
[RED] ANSWER heading and answer section
[PURPLE] RATIONALE heading and rationale section
================================================================================
EXAM INSTRUCTIONS
================================================================================
1. Read each question carefully before selecting your answer.
2. For multiple-choice items, choose the single best answer.
3. For select-all-that-apply items, choose ALL correct options.
4. Review the Answer and Rationale for every question.
5. Do not skip any question — all 200 items include explanations.
6. Correct answer positions are randomized throughout the exam.
================================================================================
BEGIN EXAM
================================================================================
pg. 2
, WGU D685 EXAM
WGU D685: Practical Applications of Prompt Engineering
Comprehensive Objective Assessment – 200 Questions
Exam Instructions: This examination contains 200 questions covering foundational prompt
engineering principles, advanced prompt crafting techniques, AI model capabilities and limitations,
ethical AI and bias mitigation, legal and regulatory compliance, evaluation and iteration of prompt
performance, and scenario-based problem solving. Select the best answer for each question. Each
question includes a complete answer and rationale.
SECTION I: FOUNDATIONS OF PROMPT ENGINEERING
Questions 1–35
QUESTION 1:
What is the primary goal of prompt engineering in generative AI?
A. To write longer prompts with more details
B. To maximize the likelihood of a desired model output
C. To reduce the number of model parameters
D. To train the model from scratch using custom datasets
ANSWER:
B. To maximize the likelihood of a desired model output
RATIONALE:
Prompt engineering focuses on designing inputs that guide a generative AI model to produce
accurate, relevant, and useful outputs without modifying the model's weights or architecture. The
goal is to elicit the best possible response through careful input design, not to alter the underlying
model. Option A is incorrect because longer prompts are not inherently better. Option C is incorrect
because prompt engineering does not modify model parameters. Option D is incorrect because
training models from scratch is a separate process entirely.
QUESTION 2:
Which of the following is an example of a zero-shot prompt?
A. "Translate 'Hello' to Spanish." (without providing examples)
B. "Here is an example: 'Cat → Gato'. Now translate 'Dog'."
C. "Classify this sentiment as positive or negative after reading 5 examples."
D. "Generate a story based on the following three examples of fairy tales."
ANSWER:
A. "Translate 'Hello' to Spanish." (without providing examples)
RATIONALE:
Zero-shot prompting means the model receives no examples of the task and must rely solely on its
pg. 3
, WGU D685 EXAM
pre-trained knowledge to generate a response. Options B, C, and D all provide examples to the
model, which makes them few-shot prompts rather than zero-shot prompts.
QUESTION 3:
What is the core problem with the prompt: "Where should I eat dinner?"
A. The AI lacks sufficient training data on restaurants
B. The prompt lacks location, cuisine preference, and budget constraints
C. The AI does not understand the concept of dinner
D. The prompt is too long and complex
ANSWER:
B. The prompt lacks location, cuisine preference, and budget constraints
RATIONALE:
This prompt is too vague and lacks essential context that would help the AI provide a useful
response. Without location, cuisine preferences, budget, or other constraints, the AI cannot provide
a relevant recommendation. The AI has sufficient training data on restaurants, understands the
concept of dinner, and the prompt is actually too short and underspecified rather than too long.
QUESTION 4:
Which technique best improves multi-step reasoning in large language models?
A. Role prompting
B. Chain-of-thought (CoT) prompting
C. Length penalty adjustment
D. Temperature adjustment
ANSWER:
B. Chain-of-thought (CoT) prompting
RATIONALE:
Chain-of-thought prompting encourages the model to show intermediate reasoning steps, improving
accuracy for math, logic, and complex reasoning tasks by 40-100% according to research. Role
prompting assigns a persona but does not specifically improve multi-step reasoning. Length penalty
and temperature adjustments affect output length and randomness respectively, but do not directly
enhance reasoning capabilities.
QUESTION 5:
What does the "temperature" parameter control in LLM generation?
A. Length of output
B. Randomness or creativity of outputs
C. Grammar correctness
D. Training data size
pg. 4
