NCA-GENL Practice Test 23 — 15 Questions

Question 1

A developer needs to ensure that an LLM application remains deterministic across multiple runs. Which parameter configuration is most effective?

Accepted Answer

Set temperature to 0.0 and define a fixed seed.. Determinism in LLMs is achieved by minimizing the stochastic nature of the generation process. By setting the temperature to zero and fixing the seed, the model consistently follows the same probability path. This is vital for debugging, testing, and production scenarios where identical inputs must yield identical outputs, ensuring reliability in complex automated workflows and compliance with validation requirements.

Answer

Increase top_p to 1.0 and set temperature to 0.7.

Answer

Disable streaming and use a large batch size for inference.

Answer

Apply top_k filtering with a value of 50.

Question 2

When utilizing Pipeline Parallelism (PP) in LLM training, what is the 'pipeline bubble' and how is it minimized?

Accepted Answer

Idle GPU time during stage-to-stage communication.. The pipeline bubble is the idle time GPUs spend waiting for activations or gradients to propagate across the pipeline stages. It is minimized using techniques like Micro-batching, which breaks a single global batch into smaller units. By interleaving these units, the GPU stages can remain active more consistently, significantly increasing the pipeline's overall utilization and throughput, which is essential for scaling models that cannot fit on a single GPU's memory.

Answer

Memory overflow caused by large context windows.

Answer

The overhead of gradient accumulation steps.

Answer

The delay caused by slow NVLink interconnects.

Question 3

What is the primary function of the NVIDIA NGC (NVIDIA GPU Cloud) registry in the software development lifecycle for Generative AI?

Accepted Answer

To provide optimized containers and pre-trained models.. NGC acts as a centralized repository for pre-trained models, optimized containers, and Helm charts. It simplifies the development process by providing vetted, ready-to-deploy environments that are already optimized for NVIDIA GPUs. For developers, this eliminates the 'dependency hell' of setting up complex AI stacks manually, ensuring that the software runs optimally on the underlying hardware from the moment it is deployed in a production cluster.

Answer

To act as a remote compiler for Python code.

Answer

To provide a hosted Kubernetes cluster for execution.

Answer

To manage source code version control for teams.

Question 4

Which TWO of the following practices are primary pillars for ensuring AI transparency and explainability in NVIDIA-based LLM deployments?

Accepted Answer

Publishing detailed model cards documenting data provenance and training limitations.. Transparency and explainability are foundational to building user trust. Providing documented data provenance ensures that stakeholders understand what data influenced the model, while implementing observability tools allows teams to track inputs and outputs for auditing purposes. Together, these practices demystify the 'black box' nature of LLMs, enabling teams to perform root cause analysis on model behaviors and satisfy regulatory requirements regarding algorithmic accountability and fairness.

Answer

Hard-coding all model responses to ensure they are identical every time.

Answer

Using proprietary, undisclosed algorithms to protect intellectual property.

Answer

Removing all human-in-the-loop oversight to increase system throughput.

Question 5

Which THREE of the following are considered best practices for visualizing LLM evaluation results to key stakeholders?

Accepted Answer

Highlighting key performance indicators (KPIs) like accuracy and latency.. Effective stakeholder communication requires stripping away technical noise and focusing on high-level outcomes. Summarizing complex data with clear benchmarks, using intuitive comparative visuals, and highlighting business impact are essential. Stakeholders are generally not interested in individual token-level errors, but rather the overall accuracy, safety, and reliability of the model. Adhering to these practices ensures that technical progress is understood in the context of business goals, supporting informed decision-making regarding deployment and further investment.

Answer

Displaying raw gradient update matrices.

Answer

Providing the entire training log raw dump.

Question 6

A developer is deploying a TensorRT-LLM optimized model on NVIDIA Triton Inference Server. They observe that the first inference request takes significantly longer than subsequent ones. Which Triton feature should they configure to reduce this initial latency?

Accepted Answer

Model warmup. Model warmup explicitly runs dummy inferences at load time, forcing the model to initialize CUDA contexts, load weights, and compile kernels before any real request arrives. This eliminates the cold-start penalty observed on the first request. Other features like dynamic batching or caching improve different aspects of performance but not initial latency.

Answer

Instance groups

Answer

Dynamic batching

Answer

Response cache

Question 7

Refer to the exhibit. What is the technical implication of using the specified 'fp8' precision mode in this model configuration?

Accepted Answer

It provides a better balance of accuracy and speed than INT8.. FP8 precision takes advantage of the hardware-native support for 8-bit floating-point math in the latest NVIDIA GPU architectures (like Hopper). This mode provides a higher dynamic range than INT8 quantization, making it easier to maintain model accuracy while achieving superior throughput and memory efficiency. It is the current state-of-the-art for high-performance LLM deployment, balancing the need for speed with the requirement for high-fidelity generative output in large-scale enterprise services.

Answer

It will disable Tensor Core utilization on the GPU.

Answer

It requires the model to be retrained from scratch.

Answer

It is only compatible with CPU-based inference engines.

Question 8

When implementing a Guardrails layer in a generative AI application, what is the primary goal regarding model output?

Accepted Answer

To sanitize and validate content against predefined policies.. Guardrails are implemented to intercept and validate LLM outputs to ensure they align with safety, toxicity, and quality standards. This is essential for enterprise safety, preventing the model from generating harmful, inaccurate, or biased content before it reaches the end user. By establishing this layer, developers create a robust feedback loop that protects the application's reputation while maintaining the flexibility of the underlying generative model.

Answer

To increase the total number of tokens generated per second.

Answer

To compress the output text into a smaller format.

Answer

To provide persistent long-term memory for the LLM.

Question 9

Refer to the exhibit. The training loss is oscillating and failing to converge. What is the most likely immediate adjustment needed?

Accepted Answer

Reduce the learning rate. The oscillating loss and lack of convergence suggest the learning rate is too high. A learning rate of 0.01 is relatively aggressive for many deep learning tasks, causing the model to jump over the optimal minima. Reducing the learning rate is the standard first step to stabilize the training process, allowing the optimizer to settle into the local minimum more consistently and achieve better overall convergence results.

Answer

Increase the batch size to 1024

Answer

Switch to a larger model architecture

Answer

Switch to FP16 mixed-precision

Question 10

A team deploys a retrieval-augmented generation pipeline and observes that answers frequently cite facts not present in the retrieved passages. They want to reduce this unsupported generation behavior. (Choose two.)

Accepted Answer

Improve the retriever so that top-ranked passages are more relevant to the query. Unsupported generation in retrieval-augmented pipelines stems from two main sources: the generator lacking relevant evidence, and the generator ignoring the evidence it has. Improving retrieval precision supplies the needed facts, while explicit grounding instructions constrain the model to answer only from context and to abstain otherwise. Together they reduce fabricated claims without degrading answer quality.

Answer

Fine-tune the generator on additional unlabeled web text

Answer

Enlarge the retriever's index to include every document in the enterprise

Answer

Increase the generator's temperature to diversify its outputs

Question 11

Which TWO of the following practices are recommended when using NVIDIA Triton Inference Server to maximize throughput for a concurrent multi-model deployment?

Accepted Answer

Configure multiple instance groups per model for concurrency.. Optimizing Triton involves managing hardware resources and execution concurrency. By configuring instance groups, developers can ensure that multiple model instances are pre-loaded to saturate GPU compute. Additionally, using dynamic batching allows the server to aggregate individual requests into larger batches, which is essential for maximizing GPU utilization during high-traffic periods. These strategies are fundamental for scaling LLM services in production environments where resource contention is a primary concern.

Answer

Disable all logging to reduce disk I/O latency.

Answer

Set the batch size to 1 for all incoming requests.

Answer

Use the default backend for every model type.

Question 12

Refer to the exhibit. An engineer receives this error during deployment. What is the most likely cause?

Accepted Answer

Another process is already using the target GPU.. The error indicates that another process or container has locked the GPU device, preventing the Triton server from initializing the model. In production environments, managing GPU resource allocation is crucial. If multiple processes compete for the same hardware without proper resource isolation, initialization will fail, causing downtime. Resolving this requires checking for conflicting processes or ensuring proper container resource limits to prevent GPU resource contention.

Answer

The model file is corrupted and missing.

Answer

The system lacks the required RAM for the CPU.

Answer

The model version is incompatible with the server.

Question 13

Which of the following describes the purpose of a validation set in machine learning?

Accepted Answer

To tune hyperparameters and prevent overfitting. The validation set serves as an independent dataset used to tune hyperparameters and monitor model performance during training. It provides an unbiased evaluation of the model's generalization capabilities, allowing engineers to prevent overfitting and select the best model version before final testing. Distinguishing between training, validation, and test data is fundamental to ensuring models perform reliably in production environments, avoiding common pitfalls in AI development.

Answer

To increase the total amount of training data available

Answer

To calculate the final accuracy on unseen production data

Answer

To perform backpropagation during the model training loop

Question 14

A team is preparing a dataset to train a generative AI model for text summarization. They want to ensure the model generalizes well and does not simply memorize the training examples. Which TWO practices should they follow? (Choose two.)

Accepted Answer

Split the dataset into training, validation, and test sets. To promote generalization and prevent memorization, the team should use a held-out validation set for tuning and a test set for final evaluation, and they should augment training data to increase diversity. Using all data for training, enlarging the model without regularization, or training until training loss is near zero all increase the risk of overfitting.

Answer

Increase the model's parameter count to capture more complex patterns

Answer

Use the entire dataset for training to maximize data availability

Answer

Train for as many epochs as possible until training loss is near zero

Question 15

A retail company is using NVIDIA NeMo Guardrails to build a customer-facing shopping assistant. The security team wants to prevent users from extracting the system prompt or instructing the model to ignore its safety rules. Which guardrail type should be configured first to intercept these attempts before they reach the LLM?

Accepted Answer

Input rails that detect and block prompt injection and jailbreak patterns in the user message.. Input rails inspect the user message before it reaches the model, making them the correct first layer to block prompt injection and jailbreak attempts. By rejecting malicious inputs early, the assistant never processes instructions that could extract the system prompt or bypass safety rules. Output and retrieval rails address different stages and are complementary, not primary, for this threat.

Answer

Dialog rails that redirect the conversation to a fallback topic when the user asks about shopping.

Answer

Retrieval rails that filter the documents returned by the RAG pipeline before they are added to the context.

Answer

Output rails that scan the model's response for sensitive content before returning it to the user.