100% Latest & Updated Google Generative AI Leader Practice Test Questions, Exam Dumps & Verified Answers!
30 Days Free Updates, Instant Download!
Question 140
A data team stores all prompt logs forever despite a shorter approved policy. Which control is neglected?
- Retention policy
- Least-privilege access
- Pseudonymization
- Data minimization
Explanation
Correct Answer
A:Retention limits should govern prompt and response logs. Retention defines how long data is held and when it should be deleted.
Incorrect Answers
B:Least privilege limits data access to identities and tasks that need it. This does not directly meet the case requirement to expire records according to approved lifecycle rules.
C:Pseudonymization replaces direct identifiers while linkability may persist with controlled supplementary information. This does not directly meet the case requirement to expire records according to approved lifecycle rules.
D:Minimization avoids collecting or providing attributes unnecessary for the task. This does not directly meet the case requirement to expire records according to approved lifecycle rules.
Question 141
Two CRM systems describe the same customer ID using different source field names, but the actual ID values are already standardized. What preparation activity aligns the different source columns to one common target definition?
- Schema mapping
- Document parsing and chunking
- Cost-aware data selection
- Data normalization
Explanation
Correct Answer
A:Mapping makes corresponding columns interoperable. Schema mapping aligns source fields to destination definitions to support reliable use.
Incorrect Answers
B:Parsing extracts and partitions source documents into sections suitable for indexing and retrieval. This does not directly meet the case requirement to align fields from different systems.
C:Cost-aware selection considers ingestion, storage, retrieval and processing costs for relevant data. This does not directly meet the case requirement to align fields from different systems.
D:Normalization standardizes units and field representations before analysis. This does not directly meet the case requirement to align fields from different systems.
Question 142
A RAG application receives thousand-page manuals and its indexer cannot locate relevant paragraphs. What preprocessing matters?
- Data normalization
- Cost-aware data selection
- Schema mapping
- Document parsing and chunking
Explanation
Correct Answer
D:Chunks make long documents retrievable in focused passages. Parsing extracts and partitions source documents into sections suitable for indexing and retrieval.
Incorrect Answers
A:Normalization standardizes units and field representations before analysis. This does not directly meet the case requirement to split long documents into searchable passages.
B:Cost-aware selection considers ingestion, storage, retrieval and processing costs for relevant data. This does not directly meet the case requirement to split long documents into searchable passages.
C:Schema mapping aligns source fields to destination definitions to support reliable use. This does not directly meet the case requirement to split long documents into searchable passages.
Question 143
A training dataset repeats the same user complaint hundreds of times after a faulty export. What fix directly addresses this?
- Cost-aware data selection
- Permission-aware indexing
- Data deduplication
- Schema mapping
Explanation
Correct Answer
C:Removing duplicate records prevents overweighting repeated examples. Deduplication detects and removes repeated items that otherwise distort training or search results.
Incorrect Answers
A:Cost-aware selection considers ingestion, storage, retrieval and processing costs for relevant data. This does not directly meet the case requirement to remove artificially repeated records.
B:Permission-aware indexing preserves document access boundaries in enterprise retrieval. This does not directly meet the case requirement to remove artificially repeated records.
D:Schema mapping aligns source fields to destination definitions to support reliable use. This does not directly meet the case requirement to remove artificially repeated records.
Question 144
A project indexes all corporate video archives even though only policy PDFs are used, driving large bills. Which assessment is needed?
- Cost-aware data selection
- Schema mapping
- Document parsing and chunking
- Data normalization
Explanation
Correct Answer
A:Relevant source selection can balance usefulness against expense. Cost-aware selection considers ingestion, storage, retrieval and processing costs for relevant data.
Incorrect Answers
B:Schema mapping aligns source fields to destination definitions to support reliable use. This does not directly meet the case requirement to compare data processing cost with actual business value.
C:Parsing extracts and partitions source documents into sections suitable for indexing and retrieval. This does not directly meet the case requirement to compare data processing cost with actual business value.
D:Normalization standardizes units and field representations before analysis. This does not directly meet the case requirement to compare data processing cost with actual business value.
Question 145
A company wants each search user to see only documents they are allowed to read. What index design property is crucial?
- Permission-aware indexing
- Cost-aware data selection
- Document parsing and chunking
- Data deduplication
Explanation
Correct Answer
A:Index access must respect the source authorization model. Permission-aware indexing preserves document access boundaries in enterprise retrieval.
Incorrect Answers
B:Cost-aware selection considers ingestion, storage, retrieval and processing costs for relevant data. This does not directly meet the case requirement to retain per-user document access boundaries.
C:Parsing extracts and partitions source documents into sections suitable for indexing and retrieval. This does not directly meet the case requirement to retain per-user document access boundaries.
D:Deduplication detects and removes repeated items that otherwise distort training or search results. This does not directly meet the case requirement to retain per-user document access boundaries.
Question 146
One data source records temperature in Celsius and another in Fahrenheit. What preparation step prevents bad comparisons?
- Data deduplication
- Cost-aware data selection
- Schema mapping
- Data normalization
Explanation
Correct Answer
D:Unit standardization ensures measurements are comparable. Normalization standardizes units and field representations before analysis.
Incorrect Answers
A:Deduplication detects and removes repeated items that otherwise distort training or search results. This does not directly meet the case requirement to standardize units before analysis.
B:Cost-aware selection considers ingestion, storage, retrieval and processing costs for relevant data. This does not directly meet the case requirement to standardize units before analysis.
C:Schema mapping aligns source fields to destination definitions to support reliable use. This does not directly meet the case requirement to standardize units before analysis.
Question 147
A new connector changes field names without changing their meaning. Which action prevents broken downstream mappings?
- Permission-aware indexing
- Schema mapping
- Cost-aware data selection
- Document parsing and chunking
Explanation
Correct Answer
B:Schema mapping must reflect the renamed columns. Schema mapping aligns source fields to destination definitions to support reliable use.
Incorrect Answers
A:Permission-aware indexing preserves document access boundaries in enterprise retrieval. This does not directly meet the case requirement to align changed source fields to expected inputs.
C:Cost-aware selection considers ingestion, storage, retrieval and processing costs for relevant data. This does not directly meet the case requirement to align changed source fields to expected inputs.
D:Parsing extracts and partitions source documents into sections suitable for indexing and retrieval. This does not directly meet the case requirement to align changed source fields to expected inputs.
Question 148
Retrieval from a PDF handbook returns whole chapters when one paragraph would suffice. Which preprocessing improvement helps?
- Document parsing and chunking
- Cost-aware data selection
- Data normalization
- Schema mapping
Explanation
Correct Answer
A:Finer parsing and chunking can improve passage precision. Parsing extracts and partitions source documents into sections suitable for indexing and retrieval.
Incorrect Answers
B:Cost-aware selection considers ingestion, storage, retrieval and processing costs for relevant data. This does not directly meet the case requirement to retrieve focused evidence from a large document.
C:Normalization standardizes units and field representations before analysis. This does not directly meet the case requirement to retrieve focused evidence from a large document.
D:Schema mapping aligns source fields to destination definitions to support reliable use. This does not directly meet the case requirement to retrieve focused evidence from a large document.
Question 149
Duplicate product pages dominate search results after a bulk migration. Which preparation action helps?
- Data deduplication
- Cost-aware data selection
- Permission-aware indexing
- Data normalization
Explanation
Correct Answer
A:Repeated source content can crowd relevant unique records. Deduplication detects and removes repeated items that otherwise distort training or search results.
Incorrect Answers
B:Cost-aware selection considers ingestion, storage, retrieval and processing costs for relevant data. This does not directly meet the case requirement to remove redundant indexed copies.
C:Permission-aware indexing preserves document access boundaries in enterprise retrieval. This does not directly meet the case requirement to remove redundant indexed copies.
D:Normalization standardizes units and field representations before analysis. This does not directly meet the case requirement to remove redundant indexed copies.
Question 150
A confidential document appears in enterprise search for an unauthorized intern. What indexing guarantee should be checked?
- Schema mapping
- Data normalization
- Permission-aware indexing
- Cost-aware data selection
Explanation
Correct Answer
C:Search availability must not override source read permissions. Permission-aware indexing preserves document access boundaries in enterprise retrieval.
Incorrect Answers
A:Schema mapping aligns source fields to destination definitions to support reliable use. This does not directly meet the case requirement to enforce original document access in retrieval.
B:Normalization standardizes units and field representations before analysis. This does not directly meet the case requirement to enforce original document access in retrieval.
D:Cost-aware selection considers ingestion, storage, retrieval and processing costs for relevant data. This does not directly meet the case requirement to enforce original document access in retrieval.
Generative AI Leader Practice Tests
