1Z0-1127-25 Practice Test 20 — 15 Questions

Question 1

Which TWO of the following are characteristics of decoder-only models like GPT? (Select TWO)

Accepted Answer

They use masked self-attention to prevent attending to future tokens. Decoder-only models use masked self-attention (causal) and generate tokens left-to-right. They cannot use bidirectional context because future tokens are masked.

Answer

They process input through an encoder and a decoder

Answer

They use bidirectional self-attention

Answer

They are ideal for tasks requiring full bidirectional context like NER

Question 2

Which THREE are valid ways to interact with OCI Generative AI?

Accepted Answer

OCI REST API.. OCI Generative AI exposes a comprehensive REST API that allows developers to programmatically interact with the service for tasks such as text generation, summarization, and chat. The REST API uses standard HTTP methods and JSON payloads, making it accessible from any language or tool that can send HTTP requests.

Answer

OCI Mobile App.

Answer

OCI Data Science Notebooks.

Question 3

Which component of the Transformer architecture allows each token to consider the relevance of every other token in the input sequence?

Accepted Answer

Self-attention. Self-attention computes attention scores between all pairs of tokens, enabling the model to capture dependencies across the entire sequence.

Answer

Multi-head attention

Answer

Feed-forward network

Answer

Positional encoding

Question 4

A data scientist is evaluating an LLM's performance on a summarization task. They observe that the model produces fluent summaries but often misses key information. Which TWO metrics would best capture this issue? (Select TWO.)

Accepted Answer

Human evaluation with a rubric for completeness. ROUGE-L measures recall of the longest common subsequence, capturing information coverage. Human evaluation can assess completeness. BLEU emphasizes precision and fluency. BERTScore measures semantic similarity but not directly the presence of key points. Perplexity measures model confidence, not recall.

Answer

BLEU score

Answer

Perplexity

Answer

BERTScore

Question 5

A team wants to reduce hallucinations in their LLM-powered question-answering system. Which TWO techniques are most effective?

Accepted Answer

Implementing RAG to retrieve relevant documents. RAG provides factual grounding, and reducing temperature makes outputs more deterministic, reducing fabricated details.

Answer

Switching to a smaller model

Answer

Using a larger context window

Answer

Increasing the temperature to 1.5

Question 6

Which of the following sampling strategies selects tokens based on a cumulative probability threshold from the highest probability tokens?

Accepted Answer

Top-p (nucleus) sampling. Top-p (nucleus) sampling cuts off the tail of the probability distribution where cumulative probability exceeds p, allowing dynamic vocabulary size.

Answer

Top-k sampling

Answer

Greedy decoding

Answer

Temperature sampling

Question 7

A data scientist is fine-tuning a Llama 2 7B model on a custom dataset using OCI Data Science. After training, the model generates fluent but factually incorrect statements about the new domain. Which post-training technique would BEST address this issue without retraining?

Accepted Answer

Use a retrieval-augmented generation (RAG) pipeline. RAG retrieves factual information from an external knowledge base to ground the generation, reducing hallucinations. The other options do not address factual accuracy.

Answer

Decrease the temperature to 0.1

Answer

Switch to a larger model like Llama 2 70B

Answer

Apply top-p sampling with p=0.9

Question 8

An organization requires low-latency inference for a custom fine-tuned model that will be used in a real-time application. They also need guaranteed availability and isolation from other tenants. Which infrastructure option should they choose?

Accepted Answer

Provision a Dedicated AI Cluster with model units for the fine-tuned model. Dedicated AI Clusters provide isolated, low-latency inference for custom models. Shared infrastructure is multi-tenant and may have higher latency. On-demand tokens and serving endpoints without a cluster do not offer dedicated resources.

Answer

Create an endpoint on OCI API Gateway pointing to shared inference

Answer

Use the OCI Generative AI Playground to host the model

Answer

On-demand token-based inference on shared infrastructure

Question 9

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 10

An application needs to generate embeddings for customer reviews to cluster them by sentiment. Which input type should be specified in the Embedding API call?

Accepted Answer

clustering. For clustering tasks, the 'clustering' input type optimizes the embeddings for that purpose. The other types are for different tasks.

Answer

search_document

Answer

search_query

Answer

classification

Question 11

A company wants to build a RAG-based assistant that answers queries using documents stored in OCI Object Storage. Which OCI Generative AI service should they use?

Accepted Answer

OCI Generative AI Agents. OCI Generative AI Agents provides a managed RAG service that can ingest documents from Object Storage and answer questions based on the ingested knowledge base.

Answer

OCI Generative AI Playground

Answer

OCI Generative AI Embedding API

Answer

OCI Generative AI Chat API

Question 12

A data scientist is fine-tuning a model using T-Few in OCI Generative AI. They have prepared a dataset with prompt/completion pairs. Which file format is required for the training data upload?

Accepted Answer

JSONL with each line containing 'prompt' and 'completion' keys. OCI Generative AI fine-tuning expects training data in JSONL format with specific fields.

Answer

Plain text file with one prompt-completion pair per line separated by a tab

Answer

Parquet file with a 'text' column containing concatenated prompt and completion

Answer

CSV with columns 'input' and 'output'

Question 13

A data scientist is using the OCI Generative AI Playground to test a summarization model. They want to generate shorter summaries and avoid repetitive phrasing. Which parameter adjustments should they make?

Accepted Answer

Decrease temperature and increase frequency penalty. Lowering temperature reduces randomness, increasing frequency penalty penalizes repeated tokens, and setting appropriate max tokens controls output length. Stop sequences define when generation stops, and presence penalty penalizes token presence overall.

Answer

Increase temperature and increase presence penalty

Answer

Set stop sequences and disable frequency penalty

Answer

Increase temperature and decrease frequency penalty

Question 14

A company wants to reduce costs for a high-volume, latency-tolerant text generation workload using OCI Generative AI. Which TWO strategies should they consider?

Accepted Answer

Use on-demand (pay-per-token) pricing instead of a dedicated AI cluster. On-demand pricing (pay-per-token) and using a smaller/faster model (like Command R instead of R+) are cost-saving measures. Dedicated clusters increase cost; the largest model is more expensive; increasing temperature doesn't affect pricing.

Answer

Provision a dedicated AI cluster for consistent throughput

Answer

Increase the temperature parameter to reduce token usage

Answer

Use the largest available model to ensure best quality

Question 15

A team wants to use OCI Generative AI Agents to build a RAG system that answers questions from documents stored in OCI Object Storage. What must they create first?

Accepted Answer

A knowledge base that indexes the documents from Object Storage. The agent requires a knowledge base that is connected to the data source. Directly creating a session without a knowledge base and agent will fail.

Answer

A fine-tuned model on the documents

Answer

A dedicated AI cluster for the agent

Answer

An OCI Functions endpoint to process documents