3.3 Identify Grounding Techniques and Their Use Cases. GENERATIVE-AI-LEADER Practice Questions (Page 1)

Questions 1–5

1. 1foundation · easy

   What is the primary purpose of grounding in large language models?

   • ATo anchor model outputs to reliable, up-to-date information
   • BTo make the model generate more creative responses
   • CTo reduce the model's inference latency
   • DTo increase the model's training data size
   Reveal Solution Discussion Select an answer first
2. 2expert · medium

   A global customer-support chatbot must serve users in multiple countries. The team wants to minimize the risk of generating hate speech or harassment while still allowing the model to discuss sensitive topics like politics and religion in a neutral way. Which configuration approach best achieves this balance?

   • AEnable safety settings that block hate speech and harassment but allow other content categories.
   • BBlock all content categories that are not explicitly allowed.
   • CSet temperature to 0.0 to make the model deterministic.
   • DUse RAG to ground the model in a database of approved responses.
   Reveal Solution Discussion Select an answer first
4. 3foundation · easy

   What is the primary benefit of grounding with Google Search?

   • AIt allows the model to access private enterprise data
   • BIt improves the model's training efficiency
   • CIt provides access to current, web-sourced information
   • DIt eliminates the need for any external data sources
   Reveal Solution Discussion Select an answer first
5. 4application · medium

   A regional bank is building a customer-support chatbot to answer questions about its loan products. The bank's loan policies change quarterly, and the chatbot must always reflect the latest approved terms. The compliance team requires that responses never reference outdated policy documents. Which approach should the AI team take?

   • AFine-tune the base model quarterly on the latest policy documents and redeploy.
   • BUse retrieval-augmented generation (RAG) with the current policy documents as the grounding data source.
   • CSet a higher token count so the model can include more context in every response.
   • DIncrease the temperature and top-p settings so the model generates more diverse responses.
   Reveal Solution Discussion Select an answer first
6. 5application · easy

   A developer is building a feature that generates short product descriptions for an e-commerce site. The descriptions must fit within a fixed-width UI card that displays at most 300 characters. Which sampling parameter should the developer configure to enforce this constraint?

   • AMax output token count
   • BTemperature
   • CTop-p
   • DSafety settings
   Reveal Solution Discussion Select an answer first