1Z0-1127-25 Practice Test 34 — 15 Questions

Question 1

An OCI Generative AI practitioner observes that a Cohere Command model generates responses with outdated information about a recent event. The model was fine-tuned six months ago. Which technique should be applied to incorporate new knowledge without retraining the model?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Implement a RAG pipeline that indexes the latest documents into a vector store and retrieves relevant passages at query time. RAG (Retrieval-Augmented Generation) lets you inject new knowledge at query time without retraining: you index the latest documents into a vector store, retrieve the most relevant passages for the user's query, and pass them as context to the model. This keeps the base model frozen while grounding responses in current information, which is exactly the requirement.

Answer

Use a longer context window and include all new articles in the prompt

Answer

Fine-tune the model again with the new data

Answer

Increase the temperature parameter to encourage more creative outputs

Question 2

A researcher wants to compare two summarization models. Model A achieves a higher ROUGE-L score than Model B, but human evaluators prefer Model B's summaries. Which of the following is the MOST likely reason?

Accepted Answer

ROUGE-L measures n-gram overlap, which may not align with human judgment of quality. ROUGE-L measures n-gram overlap, which may not capture semantic quality. Human evaluators often prefer summaries that are fluent, coherent, and concise, even if they use different wording. The discrepancy indicates that ROUGE-L alone is insufficient for evaluation.

Answer

Model A overfits to the training data

Answer

Model B has a larger context window

Answer

Model A is an encoder-decoder model while Model B is decoder-only

Question 3

The architecture shown in the exhibit is missing a critical component for a RAG pipeline. What step is missing between receiving the user query and searching the vector store?

Accepted Answer

A query embedding step using an embedding model. In a RAG pipeline, the user query must be converted into a vector embedding using the same embedding model that was used to index the documents. Without this query embedding step, the vector store cannot perform a meaningful similarity search because it compares vectors, not raw text. Option B correctly identifies this missing transformation between receiving the query and searching the vector store.

Answer

A document chunking step

Answer

A data masking step for privacy

Answer

A reranker step after retrieval

Question 4

A developer is drafting a prompt for an OCI Generative AI model to classify customer emails into "billing", "technical", or "general". They want a stable, parseable output that downstream code can consume. Which prompting practice is MOST appropriate?

Accepted Answer

Specify the allowed category labels and instruct the model to return only one of them with no additional text.. Constraining classification output means enumerating the allowed labels and instructing the model to emit only one, with no extra text. This makes parsing deterministic and stable. Explanatory reasoning, maximum temperature, and free-form imitation all introduce variability or verbosity that breaks downstream consumption, so they are unsuitable for this classification requirement.

Answer

Ask the model to explain its reasoning in detail before stating the category.

Answer

Provide ten examples of free-form email summaries and ask the model to imitate the style.

Answer

Set temperature to its maximum so the model explores all possible categories.

Question 5

A company wants to use OCI Generative AI Agents for a question-answering system over their internal knowledge base stored in OCI Object Storage. The data consists of PDF and Word documents. What is the first step to make this data usable by the agent?

Accepted Answer

Create a knowledge base and associate it with the Object Storage bucket. To make documents in OCI Object Storage usable by a Generative AI Agent, the first step is to create a knowledge base and associate it with the Object Storage bucket. The knowledge base ingests and indexes the documents, enabling the agent to retrieve relevant content for question answering. This is the foundational step before the agent can query the data.

Answer

Use the Embedding API to generate embeddings for all documents

Answer

Create a dedicated AI cluster for inference

Answer

Create an OCI Generative AI Agent

Question 6

A practitioner wants to evaluate an LLM-generated summary against a human-written reference using a metric that focuses on recall of key information. Which metric is most appropriate?

Accepted Answer

ROUGE. ROUGE (Recall-Oriented Understudy for Gisting Evaluation) is the most appropriate metric because it specifically measures recall of key information by comparing n-gram overlap between the generated summary and a reference summary. This aligns directly with the practitioner's goal of evaluating how well the LLM-generated summary captures the essential content from the human-written reference.

Answer

BLEU

Answer

Perplexity

Answer

Cosine similarity

Question 7

Which OCI Generative AI model is specifically designed to generate embeddings for English text?

Accepted Answer

Cohere embed-english-v3.0. Cohere embed-english-v3.0 is the OCI Generative AI embedding model explicitly optimized for English-only text, producing high-quality vector representations for semantic search, clustering, and classification. It is the English counterpart to embed-multilingual-v3.0 and is the correct choice when the corpus is English and maximum English retrieval quality is desired.

Answer

Meta Llama 3 8B

Answer

Cohere embed-multilingual-v3.0

Answer

Cohere Command R

Question 8

A developer is building a multilingual search application and needs to generate embeddings for user queries in multiple languages. Which two options are correct? (Select TWO)

Accepted Answer

Use the embed-multilingual-v3.0 model. Option C is correct because embed-multilingual-v3.0 is Cohere's multilingual embedding model, explicitly trained to produce embeddings across many languages, which is required for a multilingual search application handling user queries in multiple languages. Option D is correct because when generating embeddings for a user's search query, the input type must be set to 'search_query' so the embedding is optimized for the query side of retrieval and matches the 'search_document' embeddings used for the indexed content. Option A is wrong because embed-english-v3.0 only supports English and cannot properly embed queries in other languages. Option B is wrong because 'search_document' is the input type for indexing documents, not for embedding user queries. Option E is wrong because 'classification' is an input type intended for text classification tasks, not for search query embedding.

Answer

Use the embed-english-v3.0 model

Answer

Set the input type to 'search_document'

Answer

Set the input type to 'classification'

Question 9

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

Question 10

Which TWO of the following are best practices for building a RAG pipeline in OCI?

Accepted Answer

Use overlapping chunks. Overlapping chunks ensure that context is not lost at chunk boundaries, which is critical for retrieval accuracy in RAG pipelines. By including overlapping text segments, the embedding model can capture semantic continuity, reducing the risk of missing relevant information when a query spans chunk edges.

Answer

Always use exact vector search for accuracy

Answer

Avoid storing metadata alongside vectors

Answer

Use a single large chunk for each document

Question 11

Which TWO statements about OCI Generative AI fine-tuning are true? (Choose two.)

Accepted Answer

Fine-tuning adjusts the model's weights based on custom data. Option A is correct because OCI Generative AI fine-tuning performs supervised training that updates the model's weights using the customer's custom dataset, so the model learns task-specific patterns rather than relying only on prompt context. Option E is correct because fine-tuning requires a labeled dataset of input-output pairs (for example, JSONL records with prompt and completion fields) that the service uses as training examples. Option B is wrong because fine-tuning supports datasets far larger than 10 examples, with service limits typically in the thousands of records depending on the model. Option C is wrong because fine-tuning creates a custom, dedicated fine-tuned model in the customer's tenancy and does not permanently alter the shared base model. Option D is wrong because few-shot prompting supplies examples at inference time without changing weights, whereas fine-tuning actually trains and persists new weights.

Answer

Fine-tuning can only handle up to 10 examples

Answer

Fine-tuning permanently alters the base model in OCI

Answer

Fine-tuning is equivalent to providing few-shot examples in the prompt

Question 12

An organization is concerned about bias in their LLM-powered hiring assistant. Which TWO actions are MOST effective in mitigating bias?

Accepted Answer

Implement human-in-the-loop evaluation with fairness-focused rubrics. Option D is correct because human-in-the-loop evaluation with fairness-focused rubrics directly targets bias by having reviewers assess model outputs against explicit fairness criteria, catching discriminatory patterns that automated metrics may miss and enabling corrective feedback before decisions are finalized. Option E is correct because fine-tuning on a carefully curated dataset that balances demographic representation addresses bias at its source—the training data—reducing the model's tendency to learn and reproduce skewed associations from underrepresented or overrepresented groups. Options A, B, and C do not belong: a larger context window (A) only provides more input tokens and does nothing to correct biased training or outputs; increasing temperature (B) merely adds randomness and can actually amplify inconsistent or biased responses rather than mitigate bias; and using encoder-only models like BERT (C) is a task/architecture choice for classification, not a bias-mitigation measure, since such models can still inherit and reproduce bias from their training data.

Answer

Use a larger context window to include more examples

Answer

Increase the temperature parameter to introduce more randomness

Answer

Use only encoder-only models like BERT for classification

Question 13

Which TWO actions can improve the retrieval accuracy of a RAG system? (Select two.)

Accepted Answer

Use a more accurate embedding model. Option D is correct because the embedding model determines the semantic quality of the vector representations; a more accurate embedding model produces vectors that better capture meaning, so semantically similar queries and passages land closer together in the vector space, directly raising retrieval precision. Option E is correct because enriching chunk metadata (e.g., source, section, date, entity tags) and applying strict metadata filters during retrieval narrows the candidate set to contextually relevant chunks, preventing semantically similar but off-topic passages from being returned. Option A is not inherently correct: a smaller chunk size can improve granularity but also fragments context and can hurt recall, so it is not a reliable accuracy improvement for all documents. Option B is not correct because removing stop words before embedding can strip useful contextual and syntactic signals, degrading embedding quality rather than improving it. Option C is not correct because significantly increasing topK expands the number of retrieved chunks, which tends to add noise and lower precision instead of improving accuracy.

Answer

Use a smaller chunk size for all documents

Answer

Remove stop words from documents before embedding

Answer

Increase the topK parameter significantly

Question 14

A prompt engineer wants to use the self-consistency technique to improve answer reliability. Which THREE steps are part of implementing self-consistency?

Accepted Answer

Generate multiple independent responses using chain-of-thought prompting. Self-consistency works by sampling several diverse reasoning paths and then choosing the most consistent final answer, so option B is correct because it requires generating multiple independent responses using chain-of-thought prompting rather than relying on one greedy decode. Option C is correct because a non-zero temperature is needed to introduce variation across those sampled reasoning paths; with temperature 0 the model would produce essentially identical outputs and there would be nothing to aggregate. Option D is correct because the method concludes by aggregating the final answers across the sampled responses, typically via majority vote, to select the most reliable answer. Option A is not part of the technique because self-consistency depends on multiple responses, not a single high-temperature response. Option E is not part of the technique because stop sequences merely truncate generation at a token and do not create or aggregate the multiple reasoning paths that self-consistency requires.

Answer

Use a single response with high temperature

Answer

Use a stop sequence after each reasoning step

Question 15

A developer wants to deploy a RAG application using OCI Generative AI for both embedding and text generation while minimizing costs. Which strategy is most effective?

Accepted Answer

Cache frequent queries and their embeddings. Caching frequent queries and their embeddings directly reduces the number of API calls to OCI Generative AI, which is the primary cost driver for RAG applications. By storing previously computed embeddings and generated responses, repeated or similar queries avoid redundant embedding and generation operations, leading to significant cost savings. This strategy is especially effective in scenarios with high query repetition or common user intents, as it minimizes both latency and expense without sacrificing accuracy.

Answer

Use a larger generation model

Answer

Reduce chunk size to decrease embedding calls

Answer

Use a larger embedding model for better accuracy