1Z0-1127-25 Fundamentals of Large Language Models Practice Test 7 — 15 Questions

Question 1

Which TWO factors are most likely to cause hallucinations in LLMs?

Accepted Answer

High temperature. A high temperature setting increases the randomness of token sampling, making the model more likely to generate plausible-sounding but factually incorrect or nonsensical outputs. This directly contributes to hallucinations by encouraging the model to deviate from the most probable, grounded responses.

Answer

Short context window

Answer

Excessive fine-tuning

Answer

Low top-p

Question 2

Which THREE techniques are commonly used to improve the quality of text generation?

Accepted Answer

Temperature scaling. Temperature scaling is correct because it controls the randomness of token probability distributions by dividing logits before softmax; lower temperatures (e.g., 0.1) make the model more deterministic, while higher temperatures (e.g., 1.5) increase diversity. This directly influences the quality of generated text by balancing coherence and creativity.

Answer

Greedy decoding

Answer

Random sampling

Question 3

Which TWO are advantages of using LoRA for fine-tuning?

Accepted Answer

Requires less GPU memory. LoRA (Low-Rank Adaptation) reduces GPU memory requirements because it freezes the original model weights and injects trainable low-rank matrices into specific layers. This means only a tiny fraction of parameters need gradients and optimizer states, drastically lowering memory consumption during fine-tuning compared to full fine-tuning.

Answer

Guarantees higher accuracy

Answer

Increases model size

Answer

Improves inference speed

Question 4

Refer to the exhibit. The output is very short and cuts off mid-sentence. Which parameter is most likely the cause?

Accepted Answer

Max-tokens is too low. The 'max-tokens' parameter controls the maximum number of tokens (words or subwords) the model can generate in a single response. When set too low, the model is forced to stop generating output before completing its thought, resulting in a truncated or cut-off sentence. This is the most direct cause of an output that ends abruptly mid-sentence.

Answer

Temperature is too high

Answer

Model ID incorrect

Answer

Top-p is too high

Question 5

A data scientist is using OCI Data Science to fine-tune a Cohere command model on domain-specific documents. They observe that the fine-tuned model generates repetitive text. What is the most likely cause?

Accepted Answer

The training dataset lacked diversity.. Repetitive text in fine-tuned models is a classic symptom of overfitting to a narrow or homogeneous training dataset. When the domain-specific documents lack diversity in phrasing, topics, or contexts, the model learns to latch onto the most common patterns and repeats them, rather than generalizing. This is not a hyperparameter tuning issue but a data quality issue.

Answer

The number of epochs was insufficient.

Answer

The learning rate was too high.

Answer

The batch size was too small.

Question 6

A company uses OCI Generative AI to power a chatbot for customer support. They notice that the model's responses sometimes contain factual inaccuracies. Which strategy would best reduce hallucination?

Accepted Answer

Implementing Retrieval-Augmented Generation (RAG).. Retrieval-Augmented Generation (RAG) grounds the model's responses in retrieved factual information, directly reducing hallucination. Increasing temperature increases randomness, fine-tuning on a larger corpus may not fix factual accuracy, and reducing max tokens does not affect correctness.

Answer

Increasing the temperature parameter.

Answer

Reducing the max token limit.

Answer

Fine-tuning the model on a larger general corpus.

Question 7

An architect is designing a multi-tenant application using OCI Generative AI. Each tenant has custom instructions and data. To minimize cost while maintaining isolation, which deployment approach is recommended?

Accepted Answer

Shared base model with per-tenant system prompts and retrieval.. It leverages a shared base model with per-tenant system prompts and retrieval-augmented generation (RAG) to isolate custom instructions and data without the cost of dedicated endpoints. This approach minimizes compute overhead by reusing a single model instance while maintaining logical isolation through prompt engineering and vector-based retrieval, aligning with OCI's pay-as-you-go pricing model.

Answer

Dedicated fine-tuned endpoint per tenant.

Answer

On-premises deployment of open-source models.

Answer

Single large fine-tuned model with conditional logic.

Question 8

Which two factors are essential for calculating the cost of using OCI Generative AI for text generation? (Choose two.)

Accepted Answer

Number of input tokens.. The cost of using OCI Generative AI for text generation is primarily determined by the number of input tokens (the prompt you send) and the number of output tokens (the generated response). OCI charges per token processed, making these two factors essential for cost calculation.

Answer

Model architecture (encoder-only vs decoder-only).

Answer

Number of API calls per minute.

Answer

Temperature setting.

Question 9

Which three techniques are commonly used to reduce the risk of prompt injection in LLM applications? (Choose three.)

Accepted Answer

Output filtering.. Output filtering (B) is correct because it acts as a post-processing defense that scans the LLM's generated output for malicious content, such as leaked system prompts or injected commands, before it reaches the user. This technique helps mitigate the impact of successful prompt injections by catching and neutralizing harmful outputs that bypass input controls.

Answer

Enabling prompt validation against regex patterns.

Answer

Increasing temperature.

Question 10

Which three statements about transformer architecture are correct? (Choose three.)

Accepted Answer

The softmax function is used in the attention mechanism to normalize attention scores.. The softmax function is applied to the raw attention scores (the dot products between queries and keys) to convert them into a probability distribution that sums to 1. This normalization allows the model to assign a relative weight to each token in the sequence, ensuring that the weighted sum of values is stable and interpretable.

Answer

The feed-forward network applies a different set of weights for each token position.

Answer

The encoder-decoder structure is used in GPT models.

Question 11

An OCI AI Language text classification request returns the output shown. Which conclusion is most accurate?

Accepted Answer

The text is classified as Positive with high confidence.. The output shows a sentiment label of 'Positive' with a confidence score of 0.98, indicating the model is highly confident in its classification. Option B correctly identifies this as a positive sentiment with high confidence, which is the most accurate conclusion based on the provided data.

Answer

The model is uncertain about the sentiment.

Answer

The API endpoint is misconfigured.

Answer

The --endpoint parameter is optional.

Question 12

An OCI GenAI practitioner wants to deploy a model that can generate code from natural language descriptions. Which type of model is most suitable?

Accepted Answer

GPT. GPT (Generative Pre-trained Transformer) is the most suitable model for code generation from natural language because it is an autoregressive language model optimized for text generation tasks. Unlike encoder-only models, GPT generates coherent, contextually relevant sequences of tokens, making it ideal for producing code based on descriptive prompts.

Answer

T5

Answer

ResNet

Answer

BERT

Question 13

During fine-tuning of a Cohere model on OCI Data Science, the loss curve shows a sharp spike after epoch 3. What is the most appropriate action?

Accepted Answer

Gradient clipping.. A sharp spike in the loss curve after epoch 3 during fine-tuning indicates a gradient explosion, where the gradients become excessively large and destabilize the model's weights. Gradient clipping is the most appropriate action because it directly caps the gradient norm (e.g., using `max_grad_norm=1.0` in Cohere's fine-tuning API) to prevent these spikes, ensuring stable training without altering the learning dynamics.

Answer

Reduce learning rate.

Answer

Add more training data.

Answer

Increase batch size.

Question 14

A team uses OCI Generative AI's summarization feature to condense legal documents. The summaries sometimes omit critical clauses. Which parameter adjustment is most likely to improve completeness?

Accepted Answer

Increase maxTokens.. Increasing maxTokens (option D) is the most direct way to improve completeness because it extends the maximum length of the generated summary, allowing the model to include more content from the source legal document. Critical clauses are often omitted when the token limit truncates the output before the model can cover all essential sections. This parameter controls the output length, not the style or randomness of the generation.

Answer

Adjust frequencyPenalty.

Answer

Increase temperature.

Answer

Decrease topP.

Question 15

Which OCI service provides pre-trained models for custom text classification without requiring fine-tuning?

Accepted Answer

OCI AI Language. B is correct because OCI AI Language provides pre-trained models that can perform custom text classification out-of-the-box without requiring fine-tuning. It offers built-in models for common NLP tasks like sentiment analysis, entity extraction, and text classification, allowing users to classify text into custom categories defined by their own labels without additional training.

Answer

OCI Generative AI

Answer

OCI Data Science

Answer

OCI Vision