Part I - The language-model interface

For each model request, application code supplies the input and handles the result. The generation process produces the response one token at a time. Part I explains what the model computes, what the surrounding software supplies, and how the interface joins them.

The visible interface uses text, while the model computation uses numbered tokens. The request must fit instructions, evidence, and output requirements within the available context. The application must check the returned content against the task’s requirements before using it or acting on it.

Chapter 1 separates what the model calculates from what the software must do. Chapter 2 turns a task into instructions, examples, and a response format that can be checked. Chapter 3 selects the rules, conversation history, evidence, tool descriptions, and user input included in a request, then explains what to record about that selection.

Figure 1: Part I follows text received by the application through tokens inside the model to a response checked by application code.

Chapters in this part

• 1. Language models as application components
• 2. Prompt design and structured output
• 3. Choosing and checking request context