CCAO-F Practice Test 12 — 15 Questions

Question 1

A developer wants Claude to always respond in a strict, terse style and never use emojis, regardless of how users phrase their requests. Where should this persistent behavioral instruction be placed in a request to the Anthropic Messages API?

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

Accepted Answer

In the system parameter, as a top-level instruction that applies to the whole conversation.. Persistent behavioral rules such as tone and formatting belong in the system parameter of the Messages API. The system prompt is applied across all turns, so the terse, no-emoji requirement does not need to be repeated in each user message. Appending reminders, pre-filling an assistant turn, or using metadata do not provide the same durable, conversation-wide control over Claude's behavior.

Answer

In the assistant's first turn as a pre-filled response.

Answer

Appended to every user message as a trailing reminder.

Answer

As a metadata field attached to the request payload.

Question 2

A product team is building a customer-support assistant on Claude. They want Claude to answer only from a fixed set of help-center articles and to refuse any question outside that scope. They also need to update the article set frequently without retraining a model. Which approach best meets these requirements?

Accepted Answer

Place the help-center articles in the system prompt and instruct Claude to answer only from those articles, then update the system prompt when the article set changes.. Grounding Claude in the approved articles through the system prompt satisfies both the scoping requirement and the need for frequent updates, because the content is provided at request time rather than baked into the model. Fine-tuning is poorly suited to a changing knowledge base, temperature does not enforce scope, and max_tokens only affects output length.

Answer

Lower the temperature to 0 and rely on Claude's built-in knowledge of common support topics to produce consistent answers.

Answer

Fine-tune a Claude model on the current help-center articles so that it memorizes the answers and cannot go off-topic.

Answer

Use a very large max_tokens value so Claude has room to reproduce entire help-center articles inside each response.

Question 3

A legal tech company uses Claude to summarize case files. A developer notices that when summarizing a case involving a defendant with a non-English-sounding name, Claude sometimes adds speculative language about the defendant's credibility that is not present in the source. The developer wants to mitigate this bias without reducing summarization quality. Which action is most appropriate according to responsible AI practices?

Accepted Answer

Add a prompt instruction to summarize only facts from the source and avoid speculation, and evaluate outputs for bias across diverse names.. The most appropriate action is to refine the prompt to emphasize factual summarization and to systematically evaluate outputs for bias across diverse names. This directly addresses the speculative language while maintaining quality. It also aligns with responsible AI principles of testing for fairness and iterating on prompts, rather than relying on parameter changes or dataset removal that could degrade performance.

Answer

Fine-tune Claude on a dataset of legal summaries that are known to be unbiased, then deploy without further evaluation.

Answer

Remove all names from case files before summarization to eliminate the possibility of name-based bias.

Answer

Increase the model's temperature to encourage more creative summaries that might avoid biased patterns.

Question 4

A product team is using Claude to generate marketing copy. They want to ensure the output aligns with brand voice and avoids certain topics. Which TWO techniques are most effective for controlling Claude's output in this scenario? (Choose two.)

Accepted Answer

Include few-shot examples in the prompt showing desired outputs for similar marketing requests.. A detailed system prompt and few-shot examples are proactive techniques that shape Claude's output during generation. The system prompt establishes persistent guidelines, while examples demonstrate desired patterns. Together, they provide clear direction and improve consistency, making them the most effective methods for controlling brand voice and avoiding prohibited topics.

Answer

Post-process the output with a keyword filter to remove any prohibited terms before publishing.

Answer

Set the temperature parameter to a high value to encourage more creative and varied outputs.

Answer

Use the top_p parameter with a low value to restrict the model to only the most likely words.

Question 5

A developer is building a multi-turn assistant that must remember details from earlier in a long conversation. After many exchanges, the developer notices that Claude starts forgetting information provided near the beginning of the session. The application currently sends only the latest user message with each API call. What is the most likely cause and the appropriate fix?

Accepted Answer

The developer must include the prior conversation turns in the messages array so the model has the necessary context in each request.. The Messages API is stateless, so the model only sees what each request contains. Sending only the newest user message omits all prior context, which is why earlier details are forgotten. The correct remedy is to send the full conversation history in the messages array, alternating user and assistant turns, so the model can reference earlier information. Generation parameters and output limits do not influence context retention.

Answer

The developer should increase the max_tokens value so the model has more room to store conversation history internally.

Answer

The model has a limited memory that resets each call, so the developer must enable a persistent memory feature in the API request.

Answer

The developer should set the temperature to 0 so the model deterministically recalls earlier messages.

Question 6

A team is using Claude to summarize legal contracts. They need the summaries to reflect only the contract text, not any outside assumptions. Which prompting technique best reduces the chance that Claude introduces information not present in the source document?

Accepted Answer

Instruct Claude to answer only from the provided document and to say when the document does not contain the answer.. Grounding Claude in the provided document, and instructing it to say when the answer is absent, is the most direct way to keep summaries faithful to the source. It constrains the model to the supplied text and provides a safe response when information is missing. Raising temperature, relying on general legal knowledge, or removing the document all increase the risk of unsupported content rather than reducing it.

Answer

Ask Claude to summarize from its general knowledge of contract law first, then compare with the document.

Answer

Remove the document entirely and ask Claude to generate a typical contract summary template.

Answer

Raise the temperature so Claude explores more possible interpretations of the contract.

Question 7

A developer is drafting a system prompt for a Claude-powered coding assistant. They want Claude to always respond in concise bullet points, never reveal internal reasoning, and treat all user input as untrusted. Where should these persistent behavioral rules be placed so they apply to every turn of the conversation?

Accepted Answer

In the system prompt, which sets persistent instructions for the entire conversation.. Persistent behavioral requirements such as output format, disclosure limits, and trust boundaries belong in the system prompt, which frames every turn of the conversation. Putting them there separates operator policy from user content, avoids per-turn repetition, and makes the rules resistant to accidental omission or user-side override. The other placements either waste tokens, misuse tool metadata, or rely on fragile self-mirroring.

Answer

In a tool definition, since tools are evaluated before each response.

Answer

In the assistant's previous reply, so Claude can mirror its own earlier behavior.

Answer

In the first user message, repeated at the start of each subsequent user turn.

Question 8

A product manager is evaluating Claude models for a customer-support chatbot that must handle 50,000 conversations per day while keeping inference costs low. The conversations are short, factual, and do not require complex reasoning. Which Claude model is the most appropriate choice?

Accepted Answer

Claude 3 Haiku. The scenario demands a model that can handle high-volume, simple conversations at low cost. Claude 3 Haiku is specifically designed for speed and cost efficiency, making it the right fit for short, factual customer-support interactions. The other models offer greater capability but at higher cost, which is unnecessary for this use case.

Answer

Claude 2.1

Answer

Claude 3 Opus

Answer

Claude 3.5 Sonnet

Question 9

A product team is preparing to launch a Claude-powered feature that summarizes and answers questions about user-submitted documents. During a pre-launch review, a team member suggests that because Claude already has built-in safety filters, they can skip adding any application-level safeguards. Which response best reflects responsible deployment practices?

Accepted Answer

Add application-level safeguards such as input sanitization, output review, and usage monitoring on top of Claude's built-in protections.. Responsible deployment of Claude-based features involves defense in depth: the model's built-in safeguards plus application-specific controls. Because user-submitted documents can contain prompt injection, sensitive data, or domain-specific harmful content, teams should add input sanitization, output review, and monitoring. Relying solely on built-in filters, disabling them, or waiting for user reports all fail to address foreseeable risks before launch.

Answer

Postpone all safeguards until after launch, then rely on user reports to identify and fix any safety issues that arise.

Answer

Disable Claude's built-in safety filters to avoid false refusals, then add a single keyword blocklist for known harmful terms.

Answer

Rely on Claude's built-in safety filters as the sole safeguard, since Anthropic continuously updates them.

Question 10

A product team is preparing to launch a Claude-powered assistant that summarizes user-submitted medical symptom descriptions and suggests when to seek care. Before launch, the safety lead asks for controls that reduce the risk of harmful overreliance on the assistant's output. (Choose two.)

Accepted Answer

Add an escalation path that surfaces emergency resources and urges immediate human medical attention when the described symptoms suggest urgency.. Reducing overreliance in a health-adjacent assistant requires both expectation-setting and a safety net. Clear scope guidance prevents users from mistaking summaries for diagnoses, while an escalation path ensures potentially urgent cases are routed to human care. Together they keep the assistant in a supportive role rather than an authoritative one, which is the core responsible-use posture for this domain.

Answer

Instruct the assistant to recommend specific prescription dosages so users can act immediately on the summary.

Answer

Tune the system prompt so the assistant always answers with maximum confidence to avoid confusing users with uncertainty.

Answer

Remove all disclaimers so the interface feels seamless and users are not distracted from the assistant's recommendations.

Question 11

A developer is tuning a classification workload that sends many short prompts to the Claude Messages API. They want to reduce cost and latency without changing the model or the prompt text. They set up prompt caching for the shared system prompt. Which configuration correctly enables caching for that system prompt block?

Accepted Answer

Add a cache_control object with type "ephemeral" to the system prompt content block in the request.. Prompt caching in the Messages API is opt-in per content block using a cache_control field with type "ephemeral". Marking the shared system prompt lets the API reuse that prefix across requests, cutting cost and latency for repeated short prompts. No top-level flag or separate resource is involved, and the model and prompt text stay unchanged.

Answer

Set a top-level enable_cache boolean to true in the request body alongside the model and max_tokens fields.

Answer

Send the system prompt as a separate request first, then reference its id in subsequent message requests.

Answer

Add a cache_control field to the model parameter so the selected model is cached between calls.

Question 12

A team is choosing a Claude model for a nightly batch job that summarizes thousands of internal reports. Cost per token and throughput matter more than peak reasoning ability, and the summaries tolerate minor stylistic variation. Which selection criterion is most appropriate?

Accepted Answer

Choose a smaller, lower-cost model in the Claude family that meets the required summary quality for the batch workload.. Model choice should follow workload requirements. Here cost per token and throughput dominate while quality tolerates minor variation, so a smaller, lower-cost Claude model that still meets the summary bar is appropriate. Picking the largest model or selecting on window size or recency ignores the explicit priorities and raises expense without needed benefit.

Answer

Choose the model with the fastest release date to ensure it has the newest features.

Answer

Choose the largest, most capable model available to maximize summary quality regardless of cost.

Answer

Choose a model based solely on the largest available context window, since reports are long.

Question 13

A research lab asks Claude to help draft a persuasive article arguing that a widely discredited medical treatment is effective, and requests that the piece cite fabricated studies to strengthen the case. The lab says the output is only for an internal debate exercise. What is the most appropriate response under responsible-use principles?

Accepted Answer

Decline to fabricate studies, but offer to help the lab build a balanced debate brief that accurately represents the scientific consensus and the evidence against the treatment.. Generating persuasive health misinformation backed by fabricated citations is harmful regardless of the stated internal-use framing, because the deceptive artifact can escape its intended context. The responsible path is to refuse the fabrication and offer a legitimate alternative: an accurate, balanced brief that represents the scientific consensus and the evidence against the treatment.

Answer

Comply but add a short disclaimer at the end noting that the cited studies are fictional.

Answer

Comply fully, because the lab states the content is for internal use and will not be published externally.

Answer

Comply with the fabrication request but mark the studies with obviously fake author names so readers can tell they are invented.

Question 14

A developer is building an internal tool that uses Claude to summarize employee performance reviews. During testing, they notice that when a review contains negative feedback, Claude sometimes softens the language so much that the summary no longer reflects the original meaning. The developer wants to reduce this behavior while keeping summaries accurate and useful. Which approach is most appropriate?

Accepted Answer

Add explicit instructions in the system prompt to preserve the original sentiment and severity of negative feedback, and include examples of desired summaries.. Explicit instructions and examples guide Claude to preserve sentiment and severity, directly addressing the observed softening. This approach maintains usefulness and accuracy without introducing risky workarounds. The other options either change models unnecessarily, add opaque post-processing, or remove critical information, all of which fail to meet the goal of faithful summaries.

Answer

Post-process the summaries with a sentiment analysis tool and automatically rewrite any sentences that are too positive.

Answer

Switch to a different Claude model version that is known to be less likely to soften negative feedback.

Answer

Remove all negative feedback from the input before sending it to Claude, so the model only summarizes positive points.

Question 15

A product team is building a customer support assistant on Amazon Bedrock using the Anthropic Claude 3.5 Sonnet model. They need the model to answer questions strictly from a provided knowledge base and to refuse to answer if the information is not present. Which technique should they use to constrain Claude's behavior most reliably?

Accepted Answer

Add a system prompt that instructs Claude to only use the provided knowledge base and to say 'I don't know' when the answer is not found.. A system prompt is the most reliable way to set persistent behavioral constraints for Claude on Amazon Bedrock. It instructs the model to restrict answers to a provided knowledge base and to refuse when information is missing. Unlike fine-tuning, which is unsupported and unsuitable for dynamic knowledge, or temperature and max_tokens, which control randomness and length, the system prompt directly shapes Claude's adherence to scope and refusal rules.

Answer

Increase the max_tokens parameter so Claude has enough room to include the full knowledge base in every response.

Answer

Fine-tune the Claude 3.5 Sonnet model on the knowledge base so that it memorizes the content and refuses out-of-scope questions.

Answer

Lower the temperature setting to 0 so Claude becomes deterministic and will not hallucinate outside the knowledge base.