A data engineering team is building a Retrieval Augmented Generation (RAG) pipeline that heavily relies on 'SNOWFLAKE.CORTEX.EMBED_TEXT 768' to process millions of documents daily. They need to optimize for both cost and retrieval quality. Which of the following statements are true regarding the cost and performance of 'EMBED_TEXT 768' in Snowflake? (Select all that apply)

• A. For optimal retrieval quality in RAG scenarios, text should be split into chunks of no more than 512 tokens before being passed to 'EMBED TEXT 768', even if the model supports a larger context window.
• B. The 'EMBED TEXT 768' function, regardless of the 768-dimension model used, has a fixed cost of 1.50 Credits per one million Tokens processed.
• C. The 'snowflake-arctic-embed-m-vl .5 model, used by 'EMBED TEXT 768', has a context window of 512 tokens, and texts exceeding this length are truncated before embedding.
• D. The 'EMBED_TEXT 768' function is billed based on the number of 'output tokens' generated by the embedding model, as this represents the computational complexity of the vector.
• E. To minimize costs for ' EMBED_TEXT 768 operations, it is recommended to execute queries using a smaller virtual warehouse (no larger than MEDIUM), as larger warehouses do not improve performance for these functions.

A financial institution wants to develop a Snowflake-based pipeline to process call transcripts from their customer support. The pipeline needs to perform two main tasks: first, ''summarize very lengthy technical support calls'' (up to 20,000 tokens per transcript) into concise actionable insights, and second, ''classify the sentiment'' of these calls as 'positive', 'neutral', or 'negative'. Given these requirements for integration into SQL data pipelines, which combination of Snowflake Cortex functions and prompt engineering considerations would be most appropriate?

• A. Option C
• B. Option E
• C. Option D
• D. Option A
• E. Option B

A data analyst needs to use SNOWFLAKE. CORTEX. EXTRACT_ANSWER to streamline information retrieval from various contract documents. They are new to Cortex functions and want to understand access requirements and optimal usage. Which of the following statements about using SNOWFLAKE .CORTEX. EXTRACT_ANSWER are correct?

• A. The analyst's role must be granted the SNOWFLAKE. CORTEX_USER database role to execute EXTRACT_ANSWER functions.
• B. For optimal accuracy, the source_document input should be in plain English, and the question should be specific, asking for a single value.
• C. If EXTRACT_ANSWER encounters an unresolvable issue during processing, it returns NULL instead of an error, similar to TRY_COMPLETE.
• D. It is generally recommended to process multiple documents by applying EXTRACT_ANSWER to a table column containing the document texts, allowing for efficient batch processing.
• E. EXTRACT_ANSWER is the most current and recommended function for all text extraction tasks, offering multi-label and image extraction capabilities.

A data engineer is working with Snowflake Cortex Analyst to improve its ability to answer natural language questions by precisely identifying product names for filtering. They have decided to integrate a Cortex Search Service with their semantic model to enhance literal search for the 'product_name' dimension. Which of the following configurations within the semantic model's YAML file are valid and effective for this purpose?

• A. Only specifying 'sample_valueS for the 'product_name' dimension without a entry.
• B. Including in the semantic model's 'metrics' section, referencing 'product_name'.
• C. Adding a 'cortex_search_service' entry to the 'product_name' dimension with only the 'service' field:
  
• D. Adding a entry to the 'product_name' dimension, including 'literal_column' and ensuring the search service is configured to index the physical column:
  
• E. Setting true' for the 'product_name' dimension and providing an exhaustive list of to restrict the model to only those values.

A Gen AI specialist is designing a RAG pipeline utilizing Cortex Search for an application that queries a large repository of unstructured text documents. To optimize the quality of retrieval and subsequent LLM responses, what are the critical best practices and understanding of Cortex Search's mechanisms that the specialist should consider regarding text processing and tokenization?

• A. Cortex Search operates solely on vector embeddings for semantic search; keyword-based retrieval is handled by a separate, less efficient mechanism outside the core search service.
• B. When text input exceeds an embedding model's context window, Cortex Search truncates the text for both semantic embedding and keyword-based retrieval, potentially losing critical information.
• C. Embedding models with larger context windows, such as snowflake-arctic-embed-1-v2.e-8k (8000 tokens), are always superior as they allow the RAG system to process the entire document as a single, highly relevant chunk.
• D. The SNOWFLAKE .CORTEX. COUNT TOKENS function is a helper function that can be used to accurately determine the token count for a given string based on a specified model, aiding in adherence to context window limits.
• E. For best search results, text in the search column should be split into chunks of no more than 512 tokens, as smaller chunks generally lead to more precise retrieval and relevant LLM context.