TLDR: Structure of Every LLM Chat
Date: 2026-05-13 Source: https://arpitbhayani.me/blogs/structure-of-llm-chat
Overview
Role tagging is not cosmetic. It shapes how the model responds, how context is managed across multiple turns, and how application developers constrain and direct model behaviour at a structural level. Understanding this format is the difference between using an LLM and building reliably on top of one.
Key Points
- If you have only ever interacted with a language model through a chat interface, you have seen one layer of abstraction that hides a lot of engineering.
- Why Roles Exist at All: Base language models - the kind trained purely on next-token prediction over raw text - do not have a natural concept of “conversation.” They continue text.
- The System Prompt: The system prompt is the foundational instruction layer of a conversation.
- The User Turn: The user turn is the input from the person or the system acting as a person.