Month End Sale - Special 70% Discount Offer - Ends in 0d 00h 00m 00s - Coupon code: 70dumps

NCA-GENL Questions and Answers

Question # 6

Your company has upgraded from a legacy LLM model to a new model that allows for larger sequences and higher token limits. What is the most likely result of upgrading to the new model?

A.

The number of tokens is fixed for all existing language models, so there is no benefit to upgrading to higher token limits.

B.

The newer model allows for larger context, so the outputs will improve without increasing inference time overhead.

C.

The newer model allows the same context lengths, but the larger token limit will result in more comprehensive and longer outputs with more detail.

D.

The newer model allows larger context, so outputs will improve, but you will likely incur longer inference times.

Full Access
Question # 7

Transformers are useful for language modeling because their architecture is uniquely suited for handling which of the following?

A.

Long sequences

B.

Embeddings

C.

Class tokens

D.

Translations

Full Access
Question # 8

What metrics would you use to evaluate the performance of a RAG workflow in terms of the accuracy of responses generated in relation to the input query? (Choose two.)

A.

Generator latency

B.

Retriever latency

C.

Tokens generated per second

D.

Response relevancy

E.

Context precision

Full Access
Question # 9

In the context of developing an AI application using NVIDIA’s NGC containers, how does the use of containerized environments enhance the reproducibility of LLM training and deployment workflows?

A.

Containers automatically optimize the model’s hyperparameters for better performance.

B.

Containers encapsulate dependencies and configurations, ensuring consistent execution across systems.

C.

Containers reduce the model’s memory footprint by compressing the neural network.

D.

Containers enable direct access to GPU hardware without driver installation.

Full Access
Question # 10

What are the main advantages of instructed large language models over traditional, small language models (< 300M parameters)? (Pick the 2 correct responses)

A.

Trained without the need for labeled data.

B.

Smaller latency, higher throughput.

C.

It is easier to explain the predictions.

D.

Cheaper computational costs during inference.

E.

Single generic model can do more than one task.

Full Access
Question # 11

Imagine you are training an LLM consisting of billions of parameters and your training dataset is significantly larger than the available RAM in your system. Which of the following would be an alternative?

A.

Using the GPU memory to extend the RAM capacity for storing the dataset and move the dataset in and out of the GPU, using the PCI bandwidth possibly.

B.

Using a memory-mapped file that allows the library to access and operate on elements of the dataset without needing to fully load it into memory.

C.

Discarding the excess of data and pruning the dataset to the capacity of the RAM, resulting in reduced latency during inference.

D.

Eliminating sentences that are syntactically different by semantically equivalent, possibly reducing the risk of the model hallucinating as it is trained to get to the point.

Full Access
Question # 12

Which technique is designed to train a deep learning model by adjusting the weights of the neural network based on the error between the predicted and actual outputs?

A.

Gradient Boosting

B.

Principal Component Analysis

C.

K-means Clustering

D.

Backpropagation

Full Access
Question # 13

In ML applications, which machine learning algorithm is commonly used for creating new data based on existing data?

A.

Decision tree

B.

Support vector machine

C.

Generative adversarial network

D.

K-means clustering

Full Access
Question # 14

In the context of machine learning model deployment, how can Docker be utilized to enhance the process?

A.

To automatically generate features for machine learning models.

B.

To provide a consistent environment for model training and inference.

C.

To reduce the computational resources needed for training models.

D.

To directly increase the accuracy of machine learning models.

Full Access
Question # 15

What is the correct order of steps in an ML project?

A.

Model evaluation, Data preprocessing, Model training, Data collection

B.

Model evaluation, Data collection, Data preprocessing, Model training

C.

Data preprocessing, Data collection, Model training, Model evaluation

D.

Data collection, Data preprocessing, Model training, Model evaluation

Full Access
Question # 16

Which of the following principles are widely recognized for building trustworthy AI? (Choose two.)

A.

Conversational

B.

Low latency

C.

Privacy

D.

Scalability

E.

Nondiscrimination

Full Access
Question # 17

What is a Tokenizer in Large Language Models (LLM)?

A.

A method to remove stop words and punctuation marks from text data.

B.

A machine learning algorithm that predicts the next word/token in a sequence of text.

C.

A tool used to split text into smaller units called tokens for analysis and processing.

D.

A technique used to convert text data into numerical representations called tokens for machine learning.

Full Access
Question # 18

What distinguishes BLEU scores from ROUGE scores when evaluating natural language processing models?

A.

BLEU scores determine the fluency of text generation, while ROUGE scores rate the uniqueness of generated text.

B.

BLEU scores analyze syntactic structures, while ROUGE scores evaluate semantic accuracy.

C.

BLEU scores evaluate the 'precision' of translations, while ROUGE scores focus on the 'recall' of summarized text.

D.

BLEU scores measure model efficiency, whereas ROUGE scores assess computational complexity.

Full Access
Question # 19

Which of the following best describes the purpose of attention mechanisms in transformer models?

A.

To focus on relevant parts of the input sequence for use in the downstream task.

B.

To compress the input sequence for faster processing.

C.

To generate random noise for improved model robustness.

D.

To convert text into numerical representations.

Full Access
Question # 20

In the development of Trustworthy AI, what is the significance of ‘Certification’ as a principle?

A.

It ensures that AI systems are transparent in their decision-making processes.

B.

It requires AI systems to be developed with an ethical consideration for societal impacts.

C.

It involves verifying that AI models are fit for their intended purpose according to regional or industry-specific standards.

D.

It mandates that AI models comply with relevant laws and regulations specific to their deployment region and industry.

Full Access
Question # 21

In the context of data preprocessing for Large Language Models (LLMs), what does tokenization refer to?

A.

Splitting text into smaller units like words or subwords.

B.

Converting text into numerical representations.

C.

Removing stop words from the text.

D.

Applying data augmentation techniques to generate more training data.

Full Access
Question # 22

What statement best describes the diffusion models in generative AI?

A.

Diffusion models are probabilistic generative models that progressively inject noise into data, then learn to reverse this process for sample generation.

B.

Diffusion models are discriminative models that use gradient-based optimization algorithms to classify data points.

C.

Diffusion models are unsupervised models that use clustering algorithms to group similar data points together.

D.

Diffusion models are generative models that use a transformer architecture to learn the underlying probability distribution of the data.

Full Access
Question # 23

Which of the following is an activation function used in neural networks?

A.

Sigmoid function

B.

K-means clustering function

C.

Mean Squared Error function

D.

Diffusion function

Full Access
Question # 24

When using NVIDIA RAPIDS to accelerate data preprocessing for an LLM fine-tuning pipeline, which specific feature of RAPIDS cuDF enables faster data manipulation compared to traditional CPU-based Pandas?

A.

Automatic parallelization of Python code across CPU cores.

B.

GPU-accelerated columnar data processing with zero-copy memory access.

C.

Integration with cloud-based storage for distributed data access.

D.

Conversion of Pandas DataFrames to SQL tables for faster querying.

Full Access
Question # 25

In the context of fine-tuning LLMs, which of the following metrics is most commonly used to assess the performance of a fine-tuned model?

A.

Model size

B.

Accuracy on a validation set

C.

Training duration

D.

Number of layers

Full Access
Question # 26

In the development of trustworthy AI systems, what is the primary purpose of implementing red-teaming exercises during the alignment process of large language models?

A.

To optimize the model’s inference speed for production deployment.

B.

To identify and mitigate potential biases, safety risks, and harmful outputs.

C.

To increase the model’s parameter count for better performance.

D.

To automate the collection of training data for fine-tuning.

Full Access
Question # 27

In Natural Language Processing, there are a group of steps in problem formulation collectively known as word representations (also word embeddings). Which of the following are Deep Learning models that can be used to produce these representations for NLP tasks? (Choose two.)

A.

Word2vec

B.

WordNet

C.

Kubernetes

D.

TensorRT

E.

BERT

Full Access
Question # 28

What are some methods to overcome limited throughput between CPU and GPU? (Pick the 2 correct responses)

A.

Increase the clock speed of the CPU.

B.

Using techniques like memory pooling.

C.

Upgrade the GPU to a higher-end model.

D.

Increase the number of CPU cores.

Full Access