When working with IBM Watsonx Generative AI models, it's important to configure proper stopping criteria to control when the model should terminate the text generation process. You are developing a chatbot where responses should stay within a manageable length without losing coherence.
Which configuration best represents an effective stopping criterion to ensure coherent responses without abrupt truncation?

• A. Greedy decoding with temperature set to 2.0 and no stop sequence.
• B. Greedy decoding with no stop sequence and maximum tokens set to 200.
• C. Greedy decoding with maximum tokens set to 20 and a stop sequence of "END".
• D. Beam search decoding with a stop sequence of "END" and a maximum tokens limit of 50.

You are tasked with integrating IBM watsonx with a third-party customer relationship management (CRM) system to enhance customer interactions through conversational AI.
Which of the following approaches best ensures real-time responses while minimizing the latency caused by data exchange between systems?

• A. Employ an asynchronous message queue between IBM watsonx and the CRM system to handle customer requests.
• B. Utilize watsonx's real-time streaming API with pre-configured webhooks in the CRM system for instant data retrieval and response.
• C. Use IBM watsonx API for batch processing of customer requests, triggering the CRM system via a scheduled job.
• D. Set up periodic data synchronization between the CRM system and IBM watsonx using FTP file transfers.

In a RAG system, you need to select an appropriate retriever to fetch relevant documents from a large corpus before generating an answer. You are considering different types of retrievers, including embedding-based and keyword-based retrievers.
Which of the following describes a scenario where an embedding-based retriever using a vector database is the best choice?

• A. When exact keyword matching is required, and synonyms or contextual understanding are irrelevant
• B. When documents are labeled with metadata, and only metadata needs to be searched
• C. When most of the queries consist of structured queries with precise Boolean operators and relational database-style searches
• D. When retrieval must rely on semantic similarity between a query and documents, even if the exact words in the query don't appear in the document

In a Retrieval-Augmented Generation (RAG) system, you are tasked with generating vector embeddings for a large corpus of documents. You plan to use a pre-trained transformer-based model to generate these embeddings.
What is the most important factor to consider when choosing a pre-trained model for generating embeddings in this scenario?

• A. The model should have been trained on a similar task (e.g., document retrieval), as this ensures the embeddings will be relevant for your corpus.
• B. The model's embeddings should always be fine-tuned on your specific corpus before use, as pre-trained embeddings are too general for most tasks.
• C. The model's size (in terms of parameters) should be minimized to reduce memory usage, even if it impacts embedding quality.
• D. The model should generate embeddings based on sentence-level inputs only, as document-level embeddings are always too large for effective retrieval.