AI AI Engineer: NLP and Large Language Models 1

Question 1

What is the 'attention mechanism' in transformer-based models?

Accepted Answer

A mechanism that computes weighted relationships between all positions in a sequence

Answer

A method to prune unimportant neurons

Answer

A technique for data augmentation in NLP

Answer

A regularization method for language models

Question 2

What does 'tokenization' mean in the context of NLP preprocessing?

Accepted Answer

Splitting raw text into discrete units (tokens) for model input

Answer

Converting text to audio

Answer

Encrypting text data before storage

Answer

Removing stopwords from a document

Question 3

What is 'fine-tuning' a large language model (LLM)?

Accepted Answer

Continuing training of a pretrained LLM on task-specific data to adapt its behavior

Answer

Training an LLM from scratch on domain data

Answer

Compressing an LLM using quantization

Answer

Prompting an LLM with few-shot examples

Question 4

What is 'prompt engineering' when working with LLMs?

Accepted Answer

Designing input text to guide LLM behavior without modifying model weights

Answer

Writing code to deploy LLMs

Answer

Training the LLM on a new dataset

Answer

Optimizing LLM inference speed

Question 5

What is Retrieval-Augmented Generation (RAG)?

Accepted Answer

Combining an LLM with a retrieval system to ground responses in external documents

Answer

A method for training LLMs on retrieval tasks

Answer

A technique for compressing LLM context windows

Answer

Using multiple LLMs in an ensemble

Question 6

Which evaluation metric measures the overlap between generated text and human reference text using n-gram precision?

Accepted Answer

BLEU

Answer

ROUGE

Answer

Perplexity

Answer

F1 Score