D-PEN-F-A-00 시험 - DELL EMC실제시험문제와 답 - 56문항

Question No : 1

What kind of output structure would be best for programmatically parsing responses from a model?

A.Free-form narrative
B.Poetic language
C.Structured JSON or tabular format
D.Randomized paragraph layout

정답:
Explanation:
Structured formats like JSON ensure consistent response formats that can be parsed by code. This is crucial when integrating LLM outputs into pipelines or applications.

Question No : 2

Why are delimiters like triple quotes (""") useful in prompt design?

A.They encrypt the prompt
B.They compress the input
C.They clearly separate instructions, inputs, and outputs
D.They reduce training time

정답:
Explanation:
Delimiters help isolate different parts of the prompt, such as instruction vs. data, making it easier for the model to distinguish between context and command.

Question No : 3

When constructing a basic prompt, what is the ideal first step?

A.Writing model output tokens
B.Designing a structured output parser
C.Clarifying the user’s task or intent
D.Reducing the temperature to zero

정답:
Explanation:
Prompt construction starts with understanding the task and user objective. A prompt without a clearly defined intent may lead to ambiguous or irrelevant model responses.

Question No : 4

Which two prompt types help control LLM behavior with minimal examples? (Select two)

A.Zero-shot prompts
B.Few-shot prompts
C.Multi-modal prompts
D.Prompt engineering loops

정답:
Explanation:
Zero-shot prompts provide no examples, relying only on instructions, while few-shot prompts include a small number of examples to guide the model’s behavior. Both are effective without fine-tuning.

Question No : 5

Which type of prompt is best for generating concise answers?

A.Broad narrative prompt
B.Chain-of-thought prompt
C.Closed-ended prompt
D.Multi-turn prompt

정답:
Explanation:
Closed-ended prompts, such as “Yes or No” questions, constrain the model’s response space. They are ideal when a short, focused answer is required.

Question No : 6

Why is it important to specify the output format in a prompt?

A.It helps the model calculate loss
B.It ensures consistent and parsable responses
C.It optimizes the model architecture
D.It reduces token cost

정답:
Explanation:
Defining the output format―like “Answer in JSON” or “Return only one word”―helps make responses structured, accurate, and machine-readable, which is crucial for downstream automation.

Question No : 7

Which two formatting elements help improve prompt clarity? (Select two)

A.Use of structured separators (e.g., “###”, “---”)
B.Adding emojis in the prompt
C.Using ambiguous questions
D.Bullet-point or list formatting

정답:
Explanation:
Structured formatting like separators and lists help the model understand the boundaries between sections or examples. This organization reduces confusion and increases relevance in output.

Question No : 8

What is the role of examples in a prompt?

A.To decrease model latency
B.To help the model learn task-specific behavior from the prompt
C.To force deterministic outputs
D.To bypass training datasets

정답:
Explanation:
Examples in a prompt (few-shot learning) allow the model to infer the structure and expectations of the task. They’re particularly useful for classification, summarization, and formatting tasks.

Question No : 9

Which two components help structure a well-formed prompt? (Select two)

A.Context or background information
B.Delimiter expansion
C.Temperature decay
D.Expected output format

정답:
Explanation:
Providing context and a clear output format helps guide the model toward more accurate and predictable responses. These components reduce ambiguity and improve overall prompt effectiveness.

Question No : 10

Which of the following is a core component of a basic prompt?

A.Token limit
B.Instruction or task description
C.Transformer decoder
D.Output temperature

정답:
Explanation:
A good prompt starts with a clear instruction that communicates what the model is expected to do. This instruction guides the model’s understanding and generation of relevant outputs.

Question No : 11

What is the main legal concern when prompts ask the model to impersonate a real person?

A.Latency overhead
B.Defamation or identity theft
C.Token drift
D.Lossless decoding

정답:
Explanation:
Prompting an AI to impersonate real people can lead to defamation, reputational harm, or identity theft.
This is a serious legal and ethical violation, especially without consent.

Question No : 12

Which type of legal liability could an organization face if a prompt leads to harmful AI behavior?

A.Prompt optimization cost
B.Intellectual property bonus
C.Negligence or product liability claims
D.Fair-use rewards

정답:
Explanation:
If a poorly designed prompt causes harm or delivers discriminatory outputs, organizations may face lawsuits based on negligence, misrepresentation, or product liability, especially in regulated sectors.

Question No : 13

What ethical principles should guide prompt development for AI systems? (Select two)

A.Beneficence C doing good and minimizing harm
B.Exclusivity C limiting AI access to premium users
C.Non-maleficence C avoiding harm
D.Maximal extraction C using all user data possible

정답:
Explanation:
Ethical AI design emphasizes beneficence and non-maleficence. Prompt engineers must ensure that the AI does good (beneficence) and avoids generating harmful, biased, or misleading content (non-maleficence).

Question No : 14

What does the concept of "informed consent" mean in the context of prompt data collection?

A.Informing the model of its rights
B.Getting user approval before collecting or using their data in prompts
C.Automatically logging all prompt outputs
D.Encrypting the prompts after generation

정답:
Explanation:
Informed consent ensures that individuals know and agree to how their data will be used, including in AI prompts. This is a critical requirement under GDPR and other data privacy laws.

Question No : 15

Which U.S. regulation empowers California residents to control how their personal data is collected and used?

A.FERPA
B.COPPA
C.CCPA
D.DFARS

정답:
Explanation:
The California Consumer Privacy Act (CCPA) provides rights to California residents, such as the right to opt-out of data collection. Prompt engineers must ensure their designs align with these rights.

DELL EMC D-PEN-F-A-00 시험