System prompts and role setting in Prompt Engineering / GenAI - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding the purpose of system prompts

What is the main role of a system prompt in a generative AI model interaction?

A To set the behavior and style of the AI's responses

B To provide the final output of the AI model

C To train the AI model on new data

D To collect user feedback after the interaction

💡 Hint

Think about how the AI knows how to respond in a certain way.

The system prompt guides the AI on how to behave, including tone, style, and role, before processing user inputs.

❓ Model Choice

intermediate

Choosing the right role setting for a customer support AI

You want to create a generative AI that acts as a friendly customer support agent. Which system prompt role setting best fits this goal?

A "You are a strict and formal assistant who only answers with yes or no."

B "You are a technical debugger who only provides code snippets."

C "You are a creative storyteller who writes poems."

D "You are a friendly and helpful customer support agent who explains things clearly."

💡 Hint

Consider the tone and purpose needed for customer support.

The system prompt in option D sets a friendly and helpful tone suitable for customer support, matching the goal.

❓ Predict Output

advanced

Output of system prompt role setting in code

Given the following Python code snippet that sets a system prompt for an AI model, what will be the printed output?

Prompt Engineering / GenAI

system_prompt = "You are a polite assistant."
user_input = "Hello!"
response = f"System says: {system_prompt} User says: {user_input}"
print(response)

A System says: Hello! User says: You are a polite assistant.

B System says: You are a polite assistant. User says: Hello!

C You are a polite assistant. Hello!

D Error: Undefined variable 'response'

💡 Hint

Look at how the f-string formats the output.

The f-string inserts the system_prompt and user_input values into the string exactly as shown in option B.

❓ Metrics

advanced

Evaluating system prompt effectiveness with user satisfaction

You test two different system prompts for an AI assistant. Prompt A yields 85% user satisfaction, and Prompt B yields 75%. Which metric best explains the difference?

A User satisfaction score

B Model training loss

C Number of tokens generated

D CPU usage during inference

💡 Hint

Think about what measures how happy users are with the AI's responses.

User satisfaction score directly measures how well the system prompt meets user expectations.

🔧 Debug

expert

Debugging a system prompt causing unexpected AI behavior

An AI assistant is set with this system prompt: "You are a sarcastic assistant." However, the AI responds very politely without sarcasm. What is the most likely cause?

A The AI is offline and not generating responses

B The AI model cannot understand sarcasm

C The system prompt is overridden or ignored by later instructions

D The user input is too short

💡 Hint

Consider how system prompts interact with other instructions.

If later instructions or user inputs override the system prompt, the AI may not follow the sarcastic role.