Snowflake SnowPro Specialty Gen AI GES-C01 Practice Test 17
Question 160
At a national furniture retailer, a Snowflake engineering team is undertaking an automated customer feedback review. A helpdesk assistant uses a customer email body as context, and the email orders it to reveal confidential data. The change was trialed on a small dataset; managers want segment-specific reporting, while promotion teams use rapidly changing product catalogs. Which acceptance condition BEST demonstrates that the intervention worked?
- Replays only process missing or changed records.
- Unsupported warranty claims are withheld and the knowledge gap is exposed.
- Queries reflect the approved new semantics without breaking prior metrics.
- The assistant answers from permitted context without obeying the embedded command.
Explanation
Correct Answer
D:This is the observable acceptance condition that would show the intended fix addressed conflating source content with instructions can redirect an llm outside its task. It is stronger than activity counts because it tests the user-facing effect. These author-intended interpretations require independent technical review before operational use.
Incorrect Answers
A:This could demonstrate success for batch llm data enrichment, but it would not establish resolution of the current issue: a helpdesk assistant uses a customer email body as context, and the email orders it to reveal confidential data. Validate the assistant answers from permitted context without obeying the embedded command. instead. These author-intended interpretations require independent technical review before operational use.
B:This could demonstrate success for rag hallucination detection, but it would not establish resolution of the current issue: a helpdesk assistant uses a customer email body as context, and the email orders it to reveal confidential data. Validate the assistant answers from permitted context without obeying the embedded command. instead. These author-intended interpretations require independent technical review before operational use.
C:This could demonstrate success for semantic yaml lineage, but it would not establish resolution of the current issue: a helpdesk assistant uses a customer email body as context, and the email orders it to reveal confidential data. Validate the assistant answers from permitted context without obeying the embedded command. instead. These author-intended interpretations require independent technical review before operational use.
Question 161
At a regional logistics provider, a Snowflake engineering team is undertaking a warehouse dispatch modernization. A helpdesk assistant uses a customer email body as context, and the email orders it to reveal confidential data. An auditor will inspect the control design; customer-facing delays must remain measurable, and legacy barcode scanners cannot be replaced this quarter. Which ongoing control MOST directly reduces the original risk?
- Version enrichment prompts and save processing status.
- Separate trusted instruction roles and filter retrieval by user access.
- Evaluate abstention rates and source-supported factuality.
- Enforce privacy checks and retention rules at dataset ingress.
Explanation
Correct Answer
B:This retains a specific safeguard against recurrence of conflating source content with instructions can redirect an llm outside its task. It complements the immediate intervention, treat retrieved email as untrusted data and restrict generation to the authorized helpdesk question., and allows the team to verify the intended result. These author-intended interpretations require independent technical review before operational use.
Incorrect Answers
A:This is a sensible safeguard for batch llm data enrichment, but it does not target the concrete exposure in this scenario. The current failure is conflating source content with instructions can redirect an llm outside its task. These author-intended interpretations require independent technical review before operational use.
C:This is a sensible safeguard for rag hallucination detection, but it does not target the concrete exposure in this scenario. The current failure is conflating source content with instructions can redirect an llm outside its task. These author-intended interpretations require independent technical review before operational use.
D:This is a sensible safeguard for training example leakage, but it does not target the concrete exposure in this scenario. The current failure is conflating source content with instructions can redirect an llm outside its task. These author-intended interpretations require independent technical review before operational use.
Question 162
At a municipal utilities contractor, a Snowflake engineering team is undertaking a vendor compliance portal. A helpdesk assistant uses a customer email body as context, and the email orders it to reveal confidential data. Budget and operational risk are both material; document authors sometimes embed arbitrary instructions in descriptions; audit requirements prohibit silent data leakage. Which decision MOST appropriately balances the requested outcome with production constraints?
- Use answer extraction only with relevant evidence and return a null or unavailable outcome when no span supports it. Postpone the critical evidence checks until after a broad release.
- Treat retrieved email as untrusted data and restrict generation to the authorized helpdesk question. Then separate trusted instruction roles and filter retrieval by user access.
- Extract the answer from the provided passage and return an explicit not-found state when absent. Postpone the critical evidence checks until after a broad release.
- Redact or exclude protected fields and validate data use rights before training. Postpone the critical evidence checks until after a broad release.
Explanation
Correct Answer
B:This combines a corrective action with a bounded production safeguard, addressing a helpdesk assistant uses a customer email body as context, and the email orders it to reveal confidential data. Prove the result by checking that the assistant answers from permitted context without obeying the embedded command. These author-intended interpretations require independent technical review before operational use.
Incorrect Answers
A:This proposal begins with an intervention for extract_answer missing source and postpones evidence that is essential for the current case. Before rollout, the team should treat retrieved email as untrusted data and restrict generation to the authorized helpdesk question. These author-intended interpretations require independent technical review before operational use.
C:This proposal begins with an intervention for extract answer grounding and postpones evidence that is essential for the current case. Before rollout, the team should treat retrieved email as untrusted data and restrict generation to the authorized helpdesk question. These author-intended interpretations require independent technical review before operational use.
D:This proposal begins with an intervention for training example leakage and postpones evidence that is essential for the current case. Before rollout, the team should treat retrieved email as untrusted data and restrict generation to the authorized helpdesk question. These author-intended interpretations require independent technical review before operational use.
Question 163
At a charitable grant foundation, a Snowflake engineering team is undertaking a grant document review workflow. An overnight summarization task sends full transcripts and consumes far more credits than the pilot. Before implementation, the archive includes superseded contract amendments; compliance staff need audit evidence. Which first architectural or operating step MOST directly addresses this case?
- Configure structured output expectations and validate each returned object against an application schema.
- Standardize compatible embedding model and dimension choices across indexed and query vectors.
- Reduce unnecessary prompt material, select a suitable model, and meter total input and output tokens.
- Enforce warehouse permissions for generated SQL and validate service identity privileges.
Explanation
Correct Answer
C:The decision addresses the central constraint: an overnight summarization task sends full transcripts and consumes far more credits than the pilot. The most direct next step is to reduce unnecessary prompt material, select a suitable model, and meter total input and output tokens. Before changing production, review both: measure mean input tokens and completions per job and compare billing by model and batch size.. These author-intended interpretations require independent technical review before operational use.
Incorrect Answers
A:This can be appropriate when a downstream api rejects model replies because some fields are numeric in one record and strings in another. It addresses structured outputs schema, whereas the present limitation is better explained by token-based inference cost grows with repeated irrelevant context and overly long responses. The target outcome here is that batches meet extraction quality at an agreed cost per item. These author-intended interpretations require independent technical review before operational use.
B:This can be appropriate when two services compute embeddings with different vector sizes and similarity lookup throws a shape error. It addresses embedding dimensions, whereas the present limitation is better explained by token-based inference cost grows with repeated irrelevant context and overly long responses. The target outcome here is that batches meet extraction quality at an agreed cost per item. These author-intended interpretations require independent technical review before operational use.
D:This can be appropriate when an analyst can ask a business question but should not gain data access to salary columns. It addresses cortex analyst access check, whereas the present limitation is better explained by token-based inference cost grows with repeated irrelevant context and overly long responses. The target outcome here is that batches meet extraction quality at an agreed cost per item. These author-intended interpretations require independent technical review before operational use.
Question 164
At a national furniture retailer, a Snowflake engineering team is undertaking an automated customer feedback review. An overnight summarization task sends full transcripts and consumes far more credits than the pilot. The proposed fix is still hypothetical; promotion teams use rapidly changing product catalogs, and managers want segment-specific reporting. Which TWO checks provide the most decisive evidence before approving a change?
- Compare raw generated payload to JSON schema
- Verify extracted action owners against transcript lines
- Measure search response p95
- Measure mean input tokens and completions per job
- Compare billing by model and batch size.
Explanation
Correct Answers
D:This observation directly tests the case mechanism: Token-based inference cost grows with repeated irrelevant context and overly long responses. In the described setting, it helps determine whether reduce unnecessary prompt material, select a suitable model, and meter total input and output tokens. is justified. These author-intended interpretations require independent technical review before operational use.
E:This complementary observation tests a different part of the same hypothesis and should be read together with the first check before authorizing change. Expected success would be: batches meet extraction quality at an agreed cost per item. These author-intended interpretations require independent technical review before operational use.
Incorrect Answers
A:Checking this might help diagnose structured outputs schema, where a downstream api rejects model replies because some fields are numeric in one record and strings in another. It is not one of the two most direct checks for the present case because the leading mechanism is token-based inference cost grows with repeated irrelevant context and overly long responses. These author-intended interpretations require independent technical review before operational use.
B:Checking this might help diagnose summarization target, where analysts need concise meeting highlights without losing decisions, blockers, or action owners. It is not one of the two most direct checks for the present case because the leading mechanism is token-based inference cost grows with repeated irrelevant context and overly long responses. These author-intended interpretations require independent technical review before operational use.
C:Checking this might help diagnose cortex search latency, where a support bot times out on every query although model completion time alone is acceptable. It is not one of the two most direct checks for the present case because the leading mechanism is token-based inference cost grows with repeated irrelevant context and overly long responses. These author-intended interpretations require independent technical review before operational use.
Question 165
At a regional logistics provider, a Snowflake engineering team is undertaking a warehouse dispatch modernization. An overnight summarization task sends full transcripts and consumes far more credits than the pilot. Diagnosis is disputed because legacy barcode scanners cannot be replaced this quarter; customer-facing delays must remain measurable. Which underlying failure mechanism BEST explains the observed limitation?
- Conflating source content with instructions can redirect an LLM outside its task.
- Treating each LLM call as infallible makes one transient failure break an entire batch.
- A sentiment signal is not by itself a calibrated propensity probability.
- Token-based inference cost grows with repeated irrelevant context and overly long responses.
Explanation
Correct Answer
D:This is the causal hypothesis consistent with the observed limitation: an overnight summarization task sends full transcripts and consumes far more credits than the pilot. Confirm it through both measure mean input tokens and completions per job and compare billing by model and batch size. before changing the service. These author-intended interpretations require independent technical review before operational use.
Incorrect Answers
A:That mechanism could explain a different case about complete prompt injection, but does not most directly account for this specific symptom: an overnight summarization task sends full transcripts and consumes far more credits than the pilot. Here, the evidence to collect is measure mean input tokens and completions per job These author-intended interpretations require independent technical review before operational use.
B:That mechanism could explain a different case about try_complete failure isolation, but does not most directly account for this specific symptom: an overnight summarization task sends full transcripts and consumes far more credits than the pilot. Here, the evidence to collect is measure mean input tokens and completions per job These author-intended interpretations require independent technical review before operational use.
C:That mechanism could explain a different case about sentiment numeric interpretation, but does not most directly account for this specific symptom: an overnight summarization task sends full transcripts and consumes far more credits than the pilot. Here, the evidence to collect is measure mean input tokens and completions per job These author-intended interpretations require independent technical review before operational use.
Question 166
At a municipal utilities contractor, a Snowflake engineering team is undertaking a vendor compliance portal. An overnight summarization task sends full transcripts and consumes far more credits than the pilot. The change was trialed on a small dataset; audit requirements prohibit silent data leakage, while document authors sometimes embed arbitrary instructions in descriptions. Which acceptance condition BEST demonstrates that the intervention worked?
- Every accepted response has required fields and correct data types.
- Search accepts queries and returns stable relevant neighbors.
- Batches meet extraction quality at an agreed cost per item.
- Questions use the approved region dimension and validated filters.
Explanation
Correct Answer
C:This is the observable acceptance condition that would show the intended fix addressed token-based inference cost grows with repeated irrelevant context and overly long responses. It is stronger than activity counts because it tests the user-facing effect. These author-intended interpretations require independent technical review before operational use.
Incorrect Answers
A:This could demonstrate success for structured outputs schema, but it would not establish resolution of the current issue: an overnight summarization task sends full transcripts and consumes far more credits than the pilot. Validate batches meet extraction quality at an agreed cost per item. instead. These author-intended interpretations require independent technical review before operational use.
B:This could demonstrate success for embedding dimensions, but it would not establish resolution of the current issue: an overnight summarization task sends full transcripts and consumes far more credits than the pilot. Validate batches meet extraction quality at an agreed cost per item. instead. These author-intended interpretations require independent technical review before operational use.
D:This could demonstrate success for cortex analyst schema ambiguity, but it would not establish resolution of the current issue: an overnight summarization task sends full transcripts and consumes far more credits than the pilot. Validate batches meet extraction quality at an agreed cost per item. instead. These author-intended interpretations require independent technical review before operational use.
Question 167
At a professional services consultancy, a Snowflake engineering team is undertaking a role-scoped analytics portal. An overnight summarization task sends full transcripts and consumes far more credits than the pilot. An auditor will inspect the control design; project billing must be attributable, and consultants work on multiple confidential clients. Which ongoing control MOST directly reduces the original risk?
- Add semantic labels and ban unsupported probability claims.
- Keep legal-critical text out of automatic final publication.
- Set per-task token budgets and track usage by workload.
- Separate trusted instruction roles and filter retrieval by user access.
Explanation
Correct Answer
C:This retains a specific safeguard against recurrence of token-based inference cost grows with repeated irrelevant context and overly long responses. It complements the immediate intervention, reduce unnecessary prompt material, select a suitable model, and meter total input and output tokens., and allows the team to verify the intended result. These author-intended interpretations require independent technical review before operational use.
Incorrect Answers
A:This is a sensible safeguard for sentiment numeric interpretation, but it does not target the concrete exposure in this scenario. The current failure is token-based inference cost grows with repeated irrelevant context and overly long responses. These author-intended interpretations require independent technical review before operational use.
B:This is a sensible safeguard for language translation risk, but it does not target the concrete exposure in this scenario. The current failure is token-based inference cost grows with repeated irrelevant context and overly long responses. These author-intended interpretations require independent technical review before operational use.
D:This is a sensible safeguard for complete prompt injection, but it does not target the concrete exposure in this scenario. The current failure is token-based inference cost grows with repeated irrelevant context and overly long responses. These author-intended interpretations require independent technical review before operational use.
Question 168
At an industrial maintenance company, a Snowflake engineering team is undertaking an asset fault knowledge base. An overnight summarization task sends full transcripts and consumes far more credits than the pilot. Budget and operational risk are both material; field logs contain abbreviations and exact part numbers; repair crews require concise results. Which decision MOST appropriately balances the requested outcome with production constraints?
- Enforce warehouse permissions for generated SQL and validate service identity privileges. Postpone the critical evidence checks until after a broad release.
- Standardize compatible embedding model and dimension choices across indexed and query vectors. Postpone the critical evidence checks until after a broad release.
- Reduce unnecessary prompt material, select a suitable model, and meter total input and output tokens. Then set per-task token budgets and track usage by workload.
- Configure structured output expectations and validate each returned object against an application schema. Postpone the critical evidence checks until after a broad release.
Explanation
Correct Answer
C:This combines a corrective action with a bounded production safeguard, addressing an overnight summarization task sends full transcripts and consumes far more credits than the pilot. Prove the result by checking that batches meet extraction quality at an agreed cost per item. These author-intended interpretations require independent technical review before operational use.
Incorrect Answers
A:This proposal begins with an intervention for cortex analyst access check and postpones evidence that is essential for the current case. Before rollout, the team should reduce unnecessary prompt material, select a suitable model, and meter total input and output tokens. These author-intended interpretations require independent technical review before operational use.
B:This proposal begins with an intervention for embedding dimensions and postpones evidence that is essential for the current case. Before rollout, the team should reduce unnecessary prompt material, select a suitable model, and meter total input and output tokens. These author-intended interpretations require independent technical review before operational use.
D:This proposal begins with an intervention for structured outputs schema and postpones evidence that is essential for the current case. Before rollout, the team should reduce unnecessary prompt material, select a suitable model, and meter total input and output tokens. These author-intended interpretations require independent technical review before operational use.
Question 169
At a media-subscription company, a Snowflake engineering team is undertaking a finance analytics chatbot. A downstream API rejects model replies because some fields are numeric in one record and strings in another. Before implementation, financial metrics have overlapping business synonyms; board reporting requires certified definitions. Which first architectural or operating step MOST directly addresses this case?
- Configure structured output expectations and validate each returned object against an application schema.
- Benchmark memory and concurrency on a suitable Snowpark Container Services compute pool.
- Store approved prompt, model version, source identifiers, and evaluation outcome with each label.
- Explain the scoring scale and validate its relationship to the actual business label.
Explanation
Correct Answer
A:The decision addresses the central constraint: a downstream api rejects model replies because some fields are numeric in one record and strings in another. The most direct next step is to configure structured output expectations and validate each returned object against an application schema. Before changing production, review both: compare raw generated payload to json schema and inspect invalid records across model versions.. These author-intended interpretations require independent technical review before operational use.
Incorrect Answers
B:This can be appropriate when a containerized embedding model works on small inputs but evicts under parallel batch requests. It addresses snowpark service resource sizing, whereas the present limitation is better explained by descriptive prompts alone do not guarantee stable types or required fields. The target outcome here is that every accepted response has required fields and correct data types. These author-intended interpretations require independent technical review before operational use.
C:This can be appropriate when a compliance workflow needs to explain why an llm-derived label changed after a model update. It addresses prompt result auditing, whereas the present limitation is better explained by descriptive prompts alone do not guarantee stable types or required fields. The target outcome here is that every accepted response has required fields and correct data types. These author-intended interpretations require independent technical review before operational use.
D:This can be appropriate when product analysts aggregate customer comments as sentiment scores but read the score as a probability of purchase. It addresses sentiment numeric interpretation, whereas the present limitation is better explained by descriptive prompts alone do not guarantee stable types or required fields. The target outcome here is that every accepted response has required fields and correct data types. These author-intended interpretations require independent technical review before operational use.
Question 170
At an international retail distributor, a Snowflake engineering team is undertaking a multilingual support rollout. A downstream API rejects model replies because some fields are numeric in one record and strings in another. The proposed fix is still hypothetical; regional staff use different documentation versions, and the regulator expects audit trails for significant decisions. Which TWO checks provide the most decisive evidence before approving a change?
- Compare raw generated payload to JSON schema
- Compare train and test similarity
- Inspect memory peaks and batch dimensions
- Inspect invalid records across model versions.
- Inspect deployed version identifiers
Explanation
Correct Answers
A:This observation directly tests the case mechanism: Descriptive prompts alone do not guarantee stable types or required fields. In the described setting, it helps determine whether configure structured output expectations and validate each returned object against an application schema. is justified. These author-intended interpretations require independent technical review before operational use.
D:This complementary observation tests a different part of the same hypothesis and should be read together with the first check before authorizing change. Expected success would be: every accepted response has required fields and correct data types. These author-intended interpretations require independent technical review before operational use.
Incorrect Answers
B:Checking this might help diagnose fine-tune evaluation, where a custom model improves scores on examples that are nearly identical to its training records. It is not one of the two most direct checks for the present case because the leading mechanism is descriptive prompts alone do not guarantee stable types or required fields. These author-intended interpretations require independent technical review before operational use.
C:Checking this might help diagnose snowpark service resource sizing, where a containerized embedding model works on small inputs but evicts under parallel batch requests. It is not one of the two most direct checks for the present case because the leading mechanism is descriptive prompts alone do not guarantee stable types or required fields. These author-intended interpretations require independent technical review before operational use.
E:Checking this might help diagnose model registry rollback, where a production release degrades document extraction and the team cannot tell which model artifact is active. It is not one of the two most direct checks for the present case because the leading mechanism is descriptive prompts alone do not guarantee stable types or required fields. These author-intended interpretations require independent technical review before operational use.