Online Access Free NCA-GENL Practice Test

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

• A.Using techniques like memory pooling.
• B.Increase the clock speed of the CPU.
• C.Increase the number of CPU cores.
• D.Upgrade the GPU to a higher-end model.

In the evaluation of Natural Language Processing (NLP) systems, what do 'validity' and 'reliability' imply regarding the selection of evaluation metrics?

• A.Validity refers to the speed of metric computation, whereas reliability pertains to the metric's performance in high-volume data processing.
• B.Validity ensures the metric accurately reflects the intended property to measure, while reliability ensures consistent results over repeated measurements.
• C.Validity involves the metric's ability to predict future trends in data, and reliability refers to its capacity to integrate with multiple data sources.
• D.Validity is concerned with the metric's computational cost, while reliability is about its applicability across different NLP platforms.

In the transformer architecture, what is the purpose of positional encoding?

• A.To remove redundant information from the input sequence.
• B.To encode the importance of each token in the input sequence.
• C.To add information about the order of each token in the input sequence.
• D.To encode the semantic meaning of each token in the input sequence.

When preprocessing text data for an LLM fine-tuning task, why is it critical to apply subword tokenization (e.
g., Byte-Pair Encoding) instead of word-based tokenization for handling rare or out-of-vocabulary words?

• A.Subword tokenization breaks words into smaller units, enabling the model to generalize to unseen words.
• B.Subword tokenization removes punctuation and special characters to simplify text input.
• C.Subword tokenization creates a fixed-size vocabulary to prevent memory overflow.
• D.Subword tokenization reduces the model's computational complexity by eliminating embeddings.

What statement best describes the diffusion models in generative AI?

• A.Diffusion models are discriminative models that use gradient-based optimization algorithms to classify data points.
• B.Diffusion models are generative models that use a transformer architecture to learn the underlying probability distribution of the data.
• C.Diffusion models are probabilistic generative models that progressively inject noise into data, then learn to reverse this process for sample generation.
• D.Diffusion models are unsupervised models that use clustering algorithms to group similar data points together.