Optimizing foundation modelsFull notesSummaryRingkasanStoriesPracticeRetrieval-Augmented Generation8 exam-style questions on this lesson.Question 1 of 8What does RAG do?ARetrains the model's weights on the company dataBCompresses the model so it runs on less hardwareCFilters toxic content out of the model outputsDAdds relevant enterprise data to the prompt as contextCheck answerQuestion 2 of 8What are vector embeddings?AVectors where related items sit close togetherBEncrypted copies of the source documentsCThe billing units used by Amazon BedrockDThe labels attached to training examplesCheck answerQuestion 3 of 8In a RAG system, where are embeddings stored for fast similarity search?AThe model's weightsBA vector databaseCAmazon S3 GlacierDA relational table with no indexesCheck answerQuestion 4 of 8. Select two.Which TWO methods does a vector database use to find the closest matches to a prompt? (Select TWO.)ACosine similarityBBLEU scoreCSMOTEDGradient descentEk-nearest neighbors (k-NN)Check answer0 of 2 selectedQuestion 5 of 8Which sequence shows the RAG workflow in order?ARetrieve, embed, generate, storeBStore, generate, embed, retrieveCEmbed, store, retrieve, generateDGenerate, retrieve, store, embedCheck answerQuestion 6 of 8. Select two.Which TWO AWS services can serve as a vector database for RAG? (Select TWO.)AAmazon PollyBAmazon TranslateCAmazon OpenSearch ServerlessDAWS CloudTrailEAmazon Aurora PostgreSQL-Compatible Edition with pgvectorCheck answer0 of 2 selectedQuestion 7 of 8A team already runs Amazon RDS for PostgreSQL and wants to add vector search for RAG without adopting a new database engine. Which option fits?AAmazon Redshift SpectrumBAmazon ElastiCache with no extensionCAmazon DynamoDB AcceleratorDThe pgvector extension in Amazon RDS for PostgreSQLCheck answerQuestion 8 of 8Why does RAG produce more accurate and current answers than the base model alone?AIt grounds the answer in retrieved company dataBIt removes the context window limit entirelyCIt raises the temperature for richer answersDIt makes the model larger with new parametersCheck answerBusiness Case5 exam-style questions on this lesson.Agents7 exam-style questions on this lesson.