Model components and configurations

In this lesson

Prepare for the AI-901 exam (Microsoft Azure AI Fundamentals). In this lesson: tokens and next-token prediction, embeddings and attention, why models invent answers and how RAG grounds them, choosing a model (LLMs, SLMs, multimodal, reasoning, embeddings) in the Microsoft Foundry model catalog, deploying a model and the deployment types (Global Standard, Data Zone Standard, Standard, Provisioned, Batch), and configuration parameters (system message, temperature, top_p, max output tokens, penalties). Includes three original practice questions.