1Z0-1127-25 OCI Generative AI Service Practice Test 4 — 15 Questions

Question 1

An administrator needs to grant a group of data scientists access to use OCI Generative AI resources in a specific compartment. Which IAM policy statement should they use?

Accepted Answer

Allow group DataScientists to use generative-ai-family in compartment ABC. The 'use' verb allows access to GenAI resources. The policy should target the specific compartment.

Answer

Allow group DataScientists to manage generative-ai-family in compartment ABC

Answer

Allow group DataScientists to inspect generative-ai-family in compartment ABC

Answer

Allow group DataScientists to read generative-ai-models in tenancy

Question 2

A machine learning engineer is fine-tuning a Cohere Command R model using T-Few. They have prepared a JSONL dataset with 500 prompt-completion pairs. After submitting the fine-tuning job, they notice the model's performance on validation data is poor. Which action is MOST likely to improve performance?

Accepted Answer

Adding more high-quality training examples to the dataset. T-Few is a parameter-efficient fine-tuning method that updates only a small subset of parameters, making it highly sensitive to the quality and quantity of training data. With only 500 examples, the model likely hasn't seen enough diverse patterns to generalize well. Adding more high-quality, representative training examples directly addresses the root cause—insufficient data—by providing more signal for the few trainable parameters to learn from. This is the most reliable way to improve validation performance in low-data regimes.

Answer

Increasing the number of training epochs

Answer

Setting the temperature to 0 in the fine-tuning configuration

Answer

Switching to a larger base model like Llama 3 70B

Question 3

A developer is using the Embedding API to create embeddings for a clustering task. They want to ensure the embeddings are optimized for clustering similar documents. Which input type should they specify?

Accepted Answer

clustering. The Cohere Embedding API accepts an input_type parameter that prepends a task-specific instruction to the text before embedding. For clustering, the correct value is 'clustering', which optimizes the vector space so semantically similar documents group tightly together. Using the wrong input_type degrades cluster cohesion.

Answer

search_query

Answer

search_document

Answer

classification

Question 4

An organization needs to choose a model for a multilingual customer support chatbot that must understand and respond in five different languages. Which TWO models available in OCI Generative AI are suitable?

Accepted Answer

Meta Llama 3. Meta Llama 3 (A) is correct because it is a multilingual large language model available in OCI Generative AI that supports text generation and conversational tasks across multiple languages, making it suitable for a multilingual support chatbot. Cohere Command R+ (D) is also correct because it is a multilingual generative LLM in OCI Generative AI designed for conversational and RAG-style workloads, so it can understand and respond in five languages. Cohere Command R (B) is not marked correct here, and Cohere Rerank (C) is a reranking model rather than a generative chatbot model, so it cannot generate responses. Cohere embed-english-v3.0 (E) is an English-only embedding model used for vector representations, not multilingual response generation.

Answer

Cohere Command R

Answer

Cohere Rerank

Answer

Cohere embed-english-v3.0

Question 5

A company wants to build a customer service chatbot that answers questions about their internal policy documents. The documents are updated monthly, and the team cannot afford to retrain a model each time. Which approach is MOST appropriate?

Accepted Answer

Use Retrieval-Augmented Generation (RAG) with the policy documents indexed in a vector store. Retrieval-Augmented Generation (RAG) separates knowledge storage from the model, indexing policy documents in a vector store and retrieving relevant chunks at query time to ground the LLM's response. This means monthly document updates only require re-indexing the changed documents, not retraining or fine-tuning the model, which directly addresses the cost and freshness constraint. RAG also reduces hallucination by grounding answers in retrieved source content and allows citations back to the policy documents.

Answer

Use a larger foundation model with a longer context window and paste all documents into each prompt

Answer

Train a custom model from scratch on the policy documents each month

Answer

Fine-tune a base LLM on the policy documents monthly

Question 6

A solutions architect is deploying a chat application that calls an OCI Generative AI endpoint hosted in the Frankfurt region. The application runs on OCI Compute instances in the Ashburn region and cannot be migrated. The architect must minimize inference latency without changing the application's code or the model. Which approach should the architect use?

Accepted Answer

Call the on-demand model endpoint in the Ashburn region instead of Frankfurt, keeping the same model and API-compatible request format.. Latency in generative AI inference is dominated by network distance when the model is hosted in another region. OCI Generative AI provides on-demand endpoints in multiple regions, so the architect can point the Ashburn compute at an Ashburn endpoint for the same model. The API surface and model identity remain the same, so no code changes are needed, and the cross-region hop is eliminated.

Answer

Enable a private endpoint on the Frankfurt Generative AI endpoint and connect to it from Ashburn over a FastConnect private virtual circuit.

Answer

Increase the max tokens parameter on each request so fewer round trips are needed to complete a conversation.

Answer

Create a dedicated AI cluster in the Ashburn region and deploy the same model on demand there.

Question 7

A platform team is building a summarization service on OCI Generative AI. They must ensure that request payloads and generated text are never stored or logged by the service, and that the inference traffic stays private and does not traverse the public internet. Which TWO actions should the team take? (Choose two.)

Accepted Answer

Create a private endpoint for the Generative AI model and access it from the application's VCN.. Two independent controls are needed: a private endpoint keeps inference traffic on the OCI network, and the service's data-retention setting prevents prompts and completions from being stored. Together they satisfy the privacy and no-storage requirements. Content moderation, client-side encryption, and NAT gateways address different concerns and do not achieve either goal.

Answer

Route requests through an OCI NAT gateway so the source IP is masked.

Answer

Disable the model's content moderation filter so prompts and responses are not inspected.

Answer

Encrypt the payload with a customer-managed key before sending it to the model endpoint.

Question 8

A platform team is deploying an OCI Generative AI dedicated AI cluster for a latency-sensitive inference workload. They must ensure the cluster can host the chosen model and that only approved applications can consume it. Which TWO actions are required? (Choose two.)

Accepted Answer

Select a hosting type and unit count that supports the model's size and throughput needs. Deploying on a dedicated AI cluster requires choosing a hosting type and unit count that can host the model and handle the intended load, then creating an inference endpoint that binds that model so applications have something to call. The other choices address retention, console experimentation, or network filtering, none of which provide the required hosting capacity or the application-level access control.

Answer

Attach a network security group to the cluster to serve as the endpoint's access control

Answer

Enable the 'Do not store model data' flag on the cluster to isolate tenants

Answer

Publish the model to the OCI Generative AI Playground for public testing

Question 9

You need to convert a set of customer support tickets into vector embeddings for a similarity search application. Which OCI Generative AI model should you use?

Accepted Answer

Cohere Embed (e.g., embed-english-v3.0). Cohere Embed models (such as embed-english-v3.0) are specifically designed to generate vector embeddings from text for semantic search, clustering, and classification. Converting support tickets into embeddings for similarity search is exactly the use case for an embedding model. OCI Generative AI offers Cohere Embed as a managed embedding service.

Answer

Cohere Rerank

Answer

Meta Llama 3

Answer

Cohere Command R

Question 10

A developer is using the OCI Generative AI Chat API with Cohere Command R+ to build a multi-turn conversational agent. They want the agent to always respond in a formal tone, regardless of the user's phrasing. Which parameter should they set in the API request to achieve this consistently?

Accepted Answer

Use the 'system' parameter (or preamble_override) to provide a system message like 'You are a formal assistant'. A system message or preamble override sets the overall behavior and tone of the assistant for the entire conversation. Temperature controls randomness; max tokens limits length; stop sequences end generation — none are suitable for defining a persistent tone.

Answer

Set the 'stop_sequences' parameter to include periods

Answer

Set the 'temperature' parameter to 0.1

Answer

Set the 'max_tokens' parameter to 200

Question 11

A company uses OCI Generative AI Agents to build a RAG application that answers questions from documents stored in OCI Object Storage. The knowledge base is updated daily. Which step is necessary to ensure the agent incorporates the latest documents?

Accepted Answer

Re-sync the knowledge base after updating the documents in Object Storage. The knowledge base indexes the data sources. To reflect changes in the source documents, you must re-sync or re-index the knowledge base. The agent endpoint or session API does not automatically refresh content.

Answer

Use the Embedding API to manually index each new document

Answer

Set the session API to refresh automatically

Answer

Recreate the agent each time documents change

Question 12

You want to test different prompts and parameters (temperature, max tokens) for a summarization task using a foundation model without writing any code. Which OCI tool should you use?

Accepted Answer

OCI Generative AI Playground. The OCI Generative AI Playground provides an interactive UI to experiment with models and parameters. CLI, SDK, and console navigation are not for interactive experimentation.

Answer

OCI Console -> Generative AI -> Models

Answer

OCI CLI with the 'oci generative-ai' commands

Answer

Python SDK with InferenceClient

Question 13

A team is fine-tuning a Llama 3 model using OCI Generative AI. The training dataset contains 10,000 prompt-completion pairs in JSONL format. After submitting the fine-tuning job, it fails with a 'Data validation error'. What is the most likely cause?

Accepted Answer

The JSONL file uses 'input' and 'output' as key names instead of 'prompt' and 'completion'. OCI fine-tuning requires a specific JSONL format with 'prompt' and 'completion' keys. If the keys are named differently (e.g., 'input'/'output'), validation fails. Model units, compartment, and base model are not typically validation issues.

Answer

The base model selected does not support fine-tuning

Answer

The dataset is in the wrong compartment

Answer

The fine-tuning job does not have enough model units allocated

Question 14

An application needs to generate embeddings for text in multiple languages (English, Spanish, French). Which OCI Generative AI embedding model should be used?

Accepted Answer

embed-multilingual-v3.0. The OCI Generative AI embedding model embed-multilingual-v3.0 is specifically designed to generate embeddings for text in multiple languages, including English, Spanish, and French. It supports over 100 languages and is optimized for cross-lingual tasks, making it the correct choice for multilingual embedding needs.

Answer

embed-english-v3.0

Answer

Cohere Command R

Answer

Meta Llama 3

Question 15

You are using the OCI Generative AI Playground with a Cohere Command R model. You want the model to generate more varied and creative responses. Which parameter should you increase?

Accepted Answer

Temperature. Temperature controls the randomness of the output. Higher temperature (e.g., 0.8) makes the model more creative and varied. Max tokens, frequency penalty, and stop sequences do not primarily affect creativity.

Answer

Stop sequences

Answer

Max tokens

Answer

Frequency penalty