Question 1

You are assisting a senior team member in developing a generative AI model to generate product descriptions. The dataset includes hundreds of features, but you notice that many features are highly correlated with each other. What should be your approach to selecting features for the model?

Options :

A : Use Principal Component Analysis (PCA) to reduce dimensionality

B : Exclude all features that are correlated

C : Select features randomly to avoid bias

D : Include all features to capture maximum information

Answer: A

Question 2

You are tasked with developing a Python script to extract features from a large corpus of text data for training a machine learning model. Under the guidance of a senior team member, which feature extraction approach should you prioritize to ensure the model captures the most relevant information?

Options :

A : Extracting and counting the frequency of individual letters in the text.

B : Calculating the average word length in each document and using it as a feature.

C : Implementing a custom script to randomly sample words from each document as features.

D : Using spaCy to extract named entities and treating them as features for the model.

Answer: D

Question 3

You are developing a chatbot using a generative AI model to assist users with technical troubleshooting. The responses generated by the model are often too verbose, overwhelming the user with unnecessary details. What prompt engineering strategy would you use to make the responses more concise and user-friendly?

Options :

A : Use multiple prompts to break down the response into smaller sections

B : Explicitly instruct the model to generate concise and to-the-point responses in the prompt

C : Set the model to a higher temperature to reduce verbosity

D : Lengthen the prompt to include more background information

Answer: B

Question 4

Which of the following best describes how a transformer-based LLM can be used to summarize a large document?

Options :

A : The model reads the entire document and creates a summary by selecting random sentences.

B : The model breaks the document into paragraphs and summarizes each individually, then combines them.

C : The model encodes the entire document into vectors, processes these through self-attention layers, and then decodes a concise summary.

D : The model matches the document against a database of pre-existing summaries and retrieves the closest match.

Answer: C

Question 5

A production system running a generative AI model is experiencing memory leaks, leading to gradual system slowdowns and crashes. Which of the following strategies would most effectively help in diagnosing and resolving this issue?

Options :

A : Increase the system-s RAM to handle more data

B : Use profiling tools to identify and optimize memory usage in the model-s code

C : Retrain the model with a focus on memory efficiency

D : Reduce the batch size used during inference

Answer: B