Cisco 810-110 - Cisco AI Technical Practitioner (AITECH) Exam
Page: 2 / 10
Total 50 questions
Question #6 (Topic: Exam A)
What happens when the input provided to an LLM exceeds its context window limit?
A. The model increases its context window size.
B. The model loses the earliest part of the text.
C. The model caches the overflow data.
D. The model switches to a different algorithm.
Answer: B
Question #7 (Topic: Exam A)
What is the purpose of a vector database in a RAG architecture?
A. to allocate resources for response generation
B. to manage the context window
C. to perform model weight updates
D. to store numerical representations of data
Answer: D
Question #8 (Topic: Exam A)
How does context window management help control the cost of a long-running AI conversation?
A. by automatically switching to a lower-cost model as the conversation history grows
B. by capping the cumulative number of input tokens sent with each new prompt
C. by increasing the context window size to reduce the total number of API calls
D. by switching to an algorithm that reduces the character-to-token ratio
Answer: B
Question #9 (Topic: Exam A)
An analyst needs a model that can analyze a network topology diagram and a set of configuration logs simultaneously. Which type of model should be selected?
A. multimodal model
B. diffusion model
C. embedding model
D. generative adversarial network
Answer: A
Question #10 (Topic: Exam A)
A practitioner is estimating the operational cost of integrating a cloud-hosted LLM API into an application. How does tokenization influence the cost of using this API?
A. Tokenization has no impact because pricing is based on a flat-rate subscription.
B. Costs are determined by the number of API calls.
C. Tokenization increases cost by requiring additional GPU run time.
D. The number of tokens determines the per-request cost.
Answer: D
