WGU D685 Practical Applications of Prompt
Engineering Objective Assessment | OA Pack | Full
Question
800 quiz with answers
1. What is the primary objective of prompt engineering in the context of large language
models?
A) To ensure the model produces outputs that are as long as possible.
B) To design and refine input queries to elicit specific, desired outputs from the model.
C) To train the model on new, unseen data to improve its general knowledge.
D) To increase the computational cost of running the model for more accurate results.
Answer: B
Rationale: Prompt engineering is fundamentally about crafting the input (the prompt) to
guide the LLM's behavior and output towards a specific goal. It is not about retraining
the model or increasing computational cost. Option A is incorrect because length is not
the primary goal; relevance and accuracy are.
2. A user asks an LLM, "Tell me about the history of Rome." The model provides a brief,
general overview. Which of the following prompt adjustments would most likely yield a
more detailed and structured response focused on the Roman Republic?
A) "Tell me about the history of Rome in great detail."
B) "Provide a detailed, chronological analysis of the political and social structures of the
Roman Republic."
C) "What is the history of Rome?"
D) "Write a long essay on Rome."
Answer: B
Rationale: Option B is the most specific. It defines the scope (Roman Republic), the
format (chronological analysis), and the subject matter (political and social structures).
Option A is vague, C is too broad, and D focuses on length, not content or specificity.
3. Which of the following best describes a "zero-shot" prompt?
A) A prompt that includes a single example of the desired output format.
B) A prompt that includes multiple examples to guide the model.
C) A prompt that asks the model to perform a task without providing any prior examples.
D) A prompt that is used to generate prompts for other models.
Answer: C
, Rationale: The "zero" in zero-shot refers to the number of examples provided. The
model is expected to use its pre-existing knowledge to perform the task based solely on
the instruction.
4. An AI practitioner is using a "few-shot" prompt to classify customer feedback as positive
or negative. What is the primary purpose of including the "shots" in the prompt?
A) To increase the number of words in the model's output.
B) To provide the model with a mathematical formula for classification.
C) To demonstrate the desired input-output mapping so the model can infer the pattern.
D) To confuse the model and test its resilience.
Answer: C
Rationale: In a few-shot prompt, the examples serve as a pattern for the model. By
seeing how the task should be performed, the model learns the format and the
reasoning, allowing it to apply the same pattern to the new, final input.
5. Which of the following is NOT a key component of a well-structured prompt?
A) Role/Perspective
B) Context/Background Information
C) The specific API key used for the model
D) Task/Goal
Answer: C
Rationale: The API key is an authentication credential, not a component of the prompt's
content. Role, context, and task are all crucial for guiding the LLM's output.
6. A prompt engineer asks the model, "Act as a seasoned financial advisor. Explain the
concept of compound interest to a 15-year-old." Which two key prompt components are
being utilized?
A) Role and Task
B) Format and Tone
C) Context and Length
D) Tone and Role
Answer: A
Rationale: The role is "seasoned financial advisor," and the task is "explain the concept
of compound interest." The audience (a 15-year-old) influences the tone, but the core
components identified here are role and task.
7. What is the main risk of an "under-specified" prompt?
A) The model will always return an error.
B) The model's output will be too creative and not factual.
C) The model will interpret the request too broadly, leading to generic or irrelevant
, outputs.
D) The model will take longer to generate a response.
Answer: C
Rationale: An under-specified prompt lacks detail or clear direction. This forces the
model to make its own assumptions, which can result in a response that doesn't align
with the user's intended goal.
8. In the context of prompt engineering, what is "temperature" a parameter of?
A) The physical temperature of the server running the AI.
B) The verbosity of the model's output.
C) The randomness or creativity of the model's output.
D) The grammatical correctness of the model's output.
Answer: C
Rationale: Temperature is a hyperparameter that controls the randomness of the
model's next-word predictions. A lower temperature makes the output more
deterministic and focused, while a higher temperature makes it more creative and
unpredictable.
9. For a task requiring high precision, such as solving a math word problem, what
temperature setting would be most appropriate?
A) 0.9
B) 1.5
C) 0.2
D) 0.0
Answer: D (or C)
Rationale: For tasks requiring factuality and precision, a low temperature (close to 0.0) is
generally preferred. This makes the model's outputs deterministic. A temperature of 0.0
will always return the most probable token, which is best for math. A value of 0.2 is also
often acceptable, providing slight flexibility but still remaining focused.
10. You are using an LLM to generate marketing slogans for a new product. Which
parameter setting would encourage the most diverse and creative range of slogans?
A) Temperature = 0.1
B) Temperature = 1.2
C) Top_p = 0.1
D) Frequency Penalty = 0.0
Answer: B
Rationale: A higher temperature (e.g., > 1.0) introduces more randomness, which is
, beneficial for brainstorming and generating creative outputs. The other options are too
restrictive (low temperature) or don't directly control creativity in the same way.
11. What is the function of the "top_p" (nucleus sampling) parameter in an LLM?
A) It sets the maximum number of tokens the model can generate.
B) It penalizes the model for using common phrases.
C) It restricts the model's token selection to the smallest set of tokens whose cumulative
probability exceeds a certain threshold (p).
D) It increases the probability of the single most likely token.
Answer: C
Rationale: Top_p sampling is a more dynamic way to control randomness. Instead of
choosing from the top k tokens (like in top_k sampling), it chooses from the smallest
group of tokens that collectively account for a certain probability mass (e.g., 0.9). This
cuts off the long tail of unlikely tokens.
12. A prompt engineer wants to ensure the model does not repeat the same phrase multiple
times in a single response. Which parameter should they adjust?
A) Temperature
B) Top_p
C) Frequency Penalty
D) Max Tokens
Answer: C
Rationale: The frequency penalty reduces the likelihood of the model repeating tokens
that have already appeared in the text. A higher frequency penalty encourages more
diverse word choice and discourages repetition.
13. Which parameter directly limits the total number of words or sub-words the model can
output?
A) Temperature
B) Top_p
C) Presence Penalty
D) Max Tokens
Answer: D
Rationale: Max Tokens (or Max Length) is a hard stop that defines the maximum length
of the model's response. This is crucial for controlling output length and managing
computational costs.
14. A developer is using an LLM API and consistently receives incomplete responses that cut
off in the middle of a sentence. Which parameter is most likely misconfigured?
A) Temperature is too high.
Engineering Objective Assessment | OA Pack | Full
Question
800 quiz with answers
1. What is the primary objective of prompt engineering in the context of large language
models?
A) To ensure the model produces outputs that are as long as possible.
B) To design and refine input queries to elicit specific, desired outputs from the model.
C) To train the model on new, unseen data to improve its general knowledge.
D) To increase the computational cost of running the model for more accurate results.
Answer: B
Rationale: Prompt engineering is fundamentally about crafting the input (the prompt) to
guide the LLM's behavior and output towards a specific goal. It is not about retraining
the model or increasing computational cost. Option A is incorrect because length is not
the primary goal; relevance and accuracy are.
2. A user asks an LLM, "Tell me about the history of Rome." The model provides a brief,
general overview. Which of the following prompt adjustments would most likely yield a
more detailed and structured response focused on the Roman Republic?
A) "Tell me about the history of Rome in great detail."
B) "Provide a detailed, chronological analysis of the political and social structures of the
Roman Republic."
C) "What is the history of Rome?"
D) "Write a long essay on Rome."
Answer: B
Rationale: Option B is the most specific. It defines the scope (Roman Republic), the
format (chronological analysis), and the subject matter (political and social structures).
Option A is vague, C is too broad, and D focuses on length, not content or specificity.
3. Which of the following best describes a "zero-shot" prompt?
A) A prompt that includes a single example of the desired output format.
B) A prompt that includes multiple examples to guide the model.
C) A prompt that asks the model to perform a task without providing any prior examples.
D) A prompt that is used to generate prompts for other models.
Answer: C
, Rationale: The "zero" in zero-shot refers to the number of examples provided. The
model is expected to use its pre-existing knowledge to perform the task based solely on
the instruction.
4. An AI practitioner is using a "few-shot" prompt to classify customer feedback as positive
or negative. What is the primary purpose of including the "shots" in the prompt?
A) To increase the number of words in the model's output.
B) To provide the model with a mathematical formula for classification.
C) To demonstrate the desired input-output mapping so the model can infer the pattern.
D) To confuse the model and test its resilience.
Answer: C
Rationale: In a few-shot prompt, the examples serve as a pattern for the model. By
seeing how the task should be performed, the model learns the format and the
reasoning, allowing it to apply the same pattern to the new, final input.
5. Which of the following is NOT a key component of a well-structured prompt?
A) Role/Perspective
B) Context/Background Information
C) The specific API key used for the model
D) Task/Goal
Answer: C
Rationale: The API key is an authentication credential, not a component of the prompt's
content. Role, context, and task are all crucial for guiding the LLM's output.
6. A prompt engineer asks the model, "Act as a seasoned financial advisor. Explain the
concept of compound interest to a 15-year-old." Which two key prompt components are
being utilized?
A) Role and Task
B) Format and Tone
C) Context and Length
D) Tone and Role
Answer: A
Rationale: The role is "seasoned financial advisor," and the task is "explain the concept
of compound interest." The audience (a 15-year-old) influences the tone, but the core
components identified here are role and task.
7. What is the main risk of an "under-specified" prompt?
A) The model will always return an error.
B) The model's output will be too creative and not factual.
C) The model will interpret the request too broadly, leading to generic or irrelevant
, outputs.
D) The model will take longer to generate a response.
Answer: C
Rationale: An under-specified prompt lacks detail or clear direction. This forces the
model to make its own assumptions, which can result in a response that doesn't align
with the user's intended goal.
8. In the context of prompt engineering, what is "temperature" a parameter of?
A) The physical temperature of the server running the AI.
B) The verbosity of the model's output.
C) The randomness or creativity of the model's output.
D) The grammatical correctness of the model's output.
Answer: C
Rationale: Temperature is a hyperparameter that controls the randomness of the
model's next-word predictions. A lower temperature makes the output more
deterministic and focused, while a higher temperature makes it more creative and
unpredictable.
9. For a task requiring high precision, such as solving a math word problem, what
temperature setting would be most appropriate?
A) 0.9
B) 1.5
C) 0.2
D) 0.0
Answer: D (or C)
Rationale: For tasks requiring factuality and precision, a low temperature (close to 0.0) is
generally preferred. This makes the model's outputs deterministic. A temperature of 0.0
will always return the most probable token, which is best for math. A value of 0.2 is also
often acceptable, providing slight flexibility but still remaining focused.
10. You are using an LLM to generate marketing slogans for a new product. Which
parameter setting would encourage the most diverse and creative range of slogans?
A) Temperature = 0.1
B) Temperature = 1.2
C) Top_p = 0.1
D) Frequency Penalty = 0.0
Answer: B
Rationale: A higher temperature (e.g., > 1.0) introduces more randomness, which is
, beneficial for brainstorming and generating creative outputs. The other options are too
restrictive (low temperature) or don't directly control creativity in the same way.
11. What is the function of the "top_p" (nucleus sampling) parameter in an LLM?
A) It sets the maximum number of tokens the model can generate.
B) It penalizes the model for using common phrases.
C) It restricts the model's token selection to the smallest set of tokens whose cumulative
probability exceeds a certain threshold (p).
D) It increases the probability of the single most likely token.
Answer: C
Rationale: Top_p sampling is a more dynamic way to control randomness. Instead of
choosing from the top k tokens (like in top_k sampling), it chooses from the smallest
group of tokens that collectively account for a certain probability mass (e.g., 0.9). This
cuts off the long tail of unlikely tokens.
12. A prompt engineer wants to ensure the model does not repeat the same phrase multiple
times in a single response. Which parameter should they adjust?
A) Temperature
B) Top_p
C) Frequency Penalty
D) Max Tokens
Answer: C
Rationale: The frequency penalty reduces the likelihood of the model repeating tokens
that have already appeared in the text. A higher frequency penalty encourages more
diverse word choice and discourages repetition.
13. Which parameter directly limits the total number of words or sub-words the model can
output?
A) Temperature
B) Top_p
C) Presence Penalty
D) Max Tokens
Answer: D
Rationale: Max Tokens (or Max Length) is a hard stop that defines the maximum length
of the model's response. This is crucial for controlling output length and managing
computational costs.
14. A developer is using an LLM API and consistently receives incomplete responses that cut
off in the middle of a sentence. Which parameter is most likely misconfigured?
A) Temperature is too high.
