You are tasked with building a generative AI model to help create automated marketing copy for a business. A key concern is the potential generation of biased or legally sensitive content, which could negatively impact the company's reputation.
Which of the following strategies would be the most effective in mitigating these model risks?

• A. Use reinforcement learning to fine-tune the model based on user feedback to eliminate bias in the long term.
• B. Include fairness metrics in the model evaluation stage to monitor for biased outputs.
• C. Use a comprehensive training dataset that includes diverse business domains to reduce biases.
• D. Implement a post-processing filter to remove any potentially offensive or legally sensitive content.

You are tasked with integrating a generative AI model on watsonx.ai into a custom business workflow. The workflow requires complex prompt chains and interaction with external APIs.
Which of the following best describes how you should approach the integration using watsonx.ai and LangChain?

• A. Write custom scripts to manage all prompt sequences manually, leveraging watsonx.ai's SDK to call the generative AI model at every step.
• B. Use only watsonx.ai's built-in APIs and SDKs for integration, as LangChain is not required for chaining multiple prompts.
• C. Implement LangChain to handle complex multi-step workflows, using watsonx.ai's APIs to generate responses at specific stages in the chain.
• D. Directly integrate the external APIs with watsonx.ai without any intermediate framework, since LangChain would add unnecessary overhead.

You are generating product descriptions for an online marketplace using a generative AI model. The output is coherent but tends to repeat the same phrases and words excessively. You decide to apply a repetition penalty to reduce this repetition while keeping the temperature set to a value that maintains creativity in the text generation.
Which of the following adjustments would best achieve this goal?

• A. Set repetition penalty to 2.0 and increase temperature from 0.7 to 1.2
• B. Set repetition penalty to 1.0 and decrease temperature from 0.8 to 0.3
• C. Set repetition penalty to 0.0
• D. Set repetition penalty to 1.5 and maintain temperature at 0.8

Your team is tasked with enhancing a document search engine using IBM watsonx Discovery. The search engine will be used to help employees quickly find relevant internal documentation, such as policy guides, project specifications, and technical manuals. You have decided to use Retrieval-Augmented Generation (RAG) to enhance this search engine's performance. The current focus is on selecting the best retriever for the task and setting up a vector database to store document embeddings.
Which of the following retriever types is most suitable for this RAG-based search engine setup, considering the diverse and unstructured nature of the documents, and why?

• A. Rule-based retriever using manually crafted rules to find documents based on predefined logic.
• B. Hybrid retriever that combines sparse retrieval methods with dense embeddings stored in a flat-file system.
• C. Sparse retriever with BM25 ranking in combination with a traditional relational database.
• D. Sparse retriever with BM25 ranking in combination with a traditional relational database.