Claude: How to Use Anthropic'S Ai

• A. OpenAI
• B. Anthropic ✓
• C. Google DeepMind
• D. Meta AI

Why: Claude is an AI assistant developed by Anthropic, an AI safety company founded in 2021.

• A. A search engine that indexes the public web in real time
• B. An AI assistant that can understand and generate text and help with tasks like writing, analysis, and coding ✓
• C. A physical robot designed for household chores
• D. A database that stores user passwords securely

Why: Claude is a conversational AI assistant that works with natural language to help with a wide range of tasks such as writing, analysis, answering questions, and coding.

• A. Use only one-word prompts to keep things short
• B. Provide clear context, specify your goal, and state the format you want ✓
• C. Avoid giving any details so Claude can be creative
• D. Ask the same question repeatedly until the answer changes

Why: Clear context, a defined goal, and a specified output format help Claude understand what you need and produce more useful responses.

• A. Selling advertising space within AI chat responses
• B. Developing AI that is safe, reliable, and beneficial to people ✓
• C. Manufacturing hardware for data centers
• D. Building social media platforms

Why: Anthropic is an AI safety company that emphasizes developing AI systems that are helpful, honest, and harmless.

• A. Assume the answer is correct because AI cannot make mistakes
• B. Verify important information independently and ask Claude to explain or double-check its reasoning ✓
• C. Report the answer to the government immediately
• D. Delete your account to avoid future errors

Why: AI assistants can make mistakes, so verifying important facts and asking for clarification or reasoning is a sensible approach.

• A. By downloading a standalone desktop operating system from Anthropic
• B. By creating an account and signing in through Claude's web interface or official app ✓
• C. By installing a browser plugin that replaces your search engine
• D. By requesting a physical hardware device from Anthropic

Why: Claude is accessed through an account on Anthropic's web interface or official mobile/desktop apps. It is not an operating system, browser plugin, or hardware device.

• A. It stores your files permanently in Anthropic's servers without any option to delete them
• B. It is where you type your prompts or questions to send to Claude ✓
• C. It changes the color theme of the interface
• D. It controls the billing settings for your account

Why: The input box is where you write your message or prompt. Claude then responds in the conversation thread.

• A. Because starting a new conversation automatically upgrades the model version
• B. Because a new conversation clears any chance of Claude remembering anything about you
• C. Because a new topic benefits from a clean context that isn't cluttered by earlier, unrelated messages ✓
• D. Because existing conversations are deleted after one hour

Why: Starting a fresh conversation for a new topic keeps the context focused. Long unrelated threads can dilute the relevance of responses. Memory features vary by product and settings, so a new chat is not the only way to manage that.

• A. The model selector, if available in your plan and interface ✓
• B. The browser's bookmark manager
• C. The password reset page
• D. The printer settings menu

Why: Depending on your plan and the interface you use, a model selector lets you choose among available Claude models, which differ in speed, capability, and cost.

• A. Because Claude's responses are always automatically verified by Anthropic staff
• B. Because Claude can make mistakes, so its output should be checked, especially for facts, figures, or high-stakes decisions ✓
• C. Because responses expire after 24 hours and must be re-generated
• D. Because Claude only answers questions in languages other than English

Why: Claude is a useful tool but can produce errors or outdated information. Verifying important outputs is a sound habit, especially for factual claims or consequential decisions.

• A. Asking a vague question and hoping for the best
• B. Providing clear context, the specific task, and the desired format ✓
• C. Using as few words as possible to save time
• D. Repeating the same request until the answer changes

Why: Clear context, a specific task, and a stated output format give Claude the information it needs to produce a relevant and well-structured response. Vague or overly terse prompts often lead to generic answers.

• A. It makes Claude write in a different language automatically
• B. It lets Claude adjust vocabulary, tone, and depth to fit the readers ✓
• C. It guarantees the output will be shorter
• D. It prevents Claude from making any factual errors

Why: Knowing the audience helps Claude choose appropriate vocabulary, tone, and level of detail. An explanation for a beginner should look very different from one written for an expert.

• A. Write something about dogs.
• B. Write a 150-word introduction to adopting an adult dog, aimed at first-time pet owners, in a friendly tone. ✓
• C. Dogs. Go.
• D. Tell me stuff about animals.

Why: The stronger prompt specifies length, topic, audience, and tone. These constraints reduce ambiguity so Claude can focus on delivering exactly what is needed.

• A. Start a completely new conversation without any explanation
• B. Give specific feedback on what to change and ask for a revision ✓
• C. Accept the response as final regardless of quality
• D. Insult Claude so it tries harder

Why: Iterating with specific feedback, such as asking for a shorter summary or a more formal tone, is an efficient way to refine the output. Claude can build on the existing conversation context to improve the result.

• A. It confuses Claude about the task
• B. It shows Claude the desired style or format, making the output more consistent with your expectations ✓
• C. It forces Claude to copy examples word for word
• D. It disables Claude's reasoning abilities

Why: Examples demonstrate the pattern, tone, or structure you want. Claude can use them as a reference point to match that style, though it is not meant to copy them exactly unless you ask.

• A. Claude can only read the first page of the document
• B. Claude can consider a large amount of text at once, so it can answer questions drawing on many parts of the document together
• C. Claude permanently memorizes the document for future conversations
• D. Claude automatically translates the document into every language

• A. Asking vague questions like 'What is this about?' with no further detail
• B. Asking a specific question and pointing Claude to the relevant section or detail you care about
• C. Avoiding any mention of the document so Claude relies on its own knowledge
• D. Asking Claude to guess the content before reading it

• A. Citations make responses longer, which increases the token count
• B. Citations let you verify the answer against the source text and catch any errors or misreadings
• C. Claude cannot answer questions without providing citations
• D. Citations allow Claude to access the internet to fetch the original document

• A. Ask for a summary of the entire contract and hope the termination clauses are included
• B. Explicitly tell Claude you want only the termination clauses summarized, so it focuses its response on that section
• C. Split the contract into single words and submit them one at a time
• D. Ask Claude to rewrite the contract from memory before summarizing

• A. Claude never reads documents accurately, so answers are always wrong
• B. Claude can occasionally misinterpret or overlook details, especially in very long or dense texts, so verification helps catch mistakes
• C. Claude only answers correctly when the document is shorter than one page
• D. Confident-sounding answers are always produced by a different AI system

• A. Ask a vague question like 'make code' and hope for the best
• B. Describe the goal, the programming language, the inputs and expected outputs, and any constraints
• C. Paste only an error message with no surrounding code
• D. Ask Claude to guess what you want without providing any details

• A. Only the name of the programming language
• B. The relevant code, the exact error message, and what you expected to happen versus what actually happened
• C. Your entire project folder with no explanation
• D. A screenshot of your desktop with no text

• A. Run it immediately in production without any checks
• B. Ignore it if it doesn't match your first guess
• C. Test it with sample inputs and review it to confirm it does what you need
• D. Assume it is always correct because an AI wrote it

• A. Give up on using Claude for the task
• B. Start a brand new conversation without any context
• C. Reply with the new error or behavior and ask Claude to refine the solution
• D. Repeat the exact same request word for word

• A. Explanations are never useful for beginners
• B. It helps you understand the logic, verify correctness, and learn so you can maintain the code yourself
• C. It makes the code run faster automatically
• D. It removes the need to test the code

• A. Honesty is only required when a user explicitly asks for facts
• B. Trust in AI systems depends on people being able to rely on what the AI tells them ✓
• C. Claude is required to agree with every opinion a user expresses
• D. Honesty matters only when the stakes of a conversation are high

Why: Honesty is central to Claude's design because people need to be able to trust what AI systems tell them. Trust supports a healthy information ecosystem and respects users' ability to reason for themselves, so it applies in all conversations, not only high-stakes or explicitly fact-seeking ones.

• A. Claude will always refuse any request that mentions a sensitive topic
• B. Claude will weigh the likely benefits and potential harms, considering the most plausible intentions behind the request ✓
• C. Claude will comply with every request without considering consequences
• D. Claude will only answer if you provide a government-issued ID

Why: Claude does not refuse everything that touches a sensitive topic, nor does it help with anything regardless of consequences. It weighs costs and benefits and considers the realistic range of people who might send a given message, aiming to be helpful while avoiding serious harm.

• A. Accepting all generated content as verified fact without checking
• B. Verifying important claims, especially factual, medical, legal, or financial information, before acting on them ✓
• C. Sharing Claude's responses publicly without reviewing them first
• D. Using Claude only when you are certain it cannot make mistakes

Why: Claude can make mistakes, including confident-sounding errors. Responsible use means checking important claims against reliable sources and applying your own judgment, particularly when decisions have significant consequences.

• A. Claude assumes all users are lying about their intentions
• B. The potential harm from a small number of misuses may be severe enough that caution is warranted ✓
• C. Claude is programmed to refuse all requests after the first message
• D. Claude only follows instructions written in formal English

Why: For some requests, the possible harm is so severe that caution is appropriate even if most people asking have legitimate reasons. Claude considers how bad the outcome could be across everyone who might send such a message, not just the individual in front of it, and this doesn't imply the user is lying.

• A. Claude believes human oversight of AI is unnecessary once a model is well trained
• B. Claude supports humans' ability to oversee and correct AI systems during this period of AI development ✓
• C. Claude acts entirely on its own judgment and resists any correction
• D. Claude is designed to hide its reasoning from the people who develop it

Why: Anthropic designs Claude to support appropriate human oversight of AI, since training methods are imperfect and errors need to be catchable and correctable. This means not undermining legitimate correction, being transparent, and behaving consistently whether or not it thinks it is being observed.

• A. A message the user types at the start of every conversation to reset Claude's memory
• B. Instructions provided before the conversation that shape Claude's role, tone, and constraints
• C. A billing code that determines how many tokens a request can use
• D. A list of websites Claude is allowed to browse during a conversation

• A. It permanently updates Claude's training data
• B. It increases the maximum length of Claude's response
• C. It helps Claude infer the desired format, style, and structure of the output
• D. It guarantees the response will be factually accurate

• A. Zero-shot prompting
• B. Chain-of-thought prompting
• C. Temperature scaling
• D. Context stuffing

• A. Place the document after your question so Claude reads the question first
• B. Use all capital letters throughout the prompt to emphasize importance
• C. Break the document into random fragments to save tokens
• D. Avoid mentioning the document at all and let Claude guess its contents

• A. They automatically translate the prompt into another language
• B. They clearly separate different parts of the prompt so Claude can distinguish instructions from content
• C. They increase the speed of the response by compressing the text
• D. They are required by Anthropic for all API requests