Prompt as State Machine

# Class 2 - The prompt as a State Machine | From Chat to Controlled LLM Workflows

> **Learn how to give your LLM a memory and control multi-step conversations.**
> 
In this class, we explore how Large Language Models (LLMs) can be modeled as state machines.  
By explicitly defining state, input, and transition logic, we can reduce behavioral variance and increase predictability in AI systems.

Understanding prompt-state architecture is fundamental for building deterministic AI workflows.

<div align="center">

[![Status](https://img.shields.io/badge/Status-Active_Course-success?style=plastic)](https://github.com/mindhack03d/SymbolicPrompting)
[![GitHub Stars](https://img.shields.io/github/stars/mindhack03d/SymbolicPrompting?style=plastic&logo=github&color=gold&cacheBus=1)](https://github.com/mindhack03d/SymbolicPrompting)
[![YouTube Playlist](https://img.shields.io/badge/YouTube-Course_Complete-success?style=plastic&logo=youtube)](https://youtube.com/playlist?list=PLNFL-2KY9QZVqoRwRzVLPN6qmDftpsjg6)
[![YouTube Playlist](https://img.shields.io/badge/YouTube-Reels_EN-success?style=plastic&logo=youtube)](https://www.youtube.com/playlist?list=PLNFL-2KY9QZXhGEfGUOrrZtzGdPESwh4l)
[![YouTube Playlist](https://img.shields.io/badge/YouTube-Reels_ES-success?style=plastic&logo=youtube)](https://youtube.com/playlist?list=PLNFL-2KY9QZUKlXC_4gnVUHoAJdd4s-AC&si=4N7ROWCD3G46y8t5l)<br>
[![License-MIT](https://img.shields.io/badge/License-MIT-green?style=plastic)](https://opensource.org/licenses/MIT)
[![Methodology-V2](https://img.shields.io/badge/Methodology-V2.0-blue?style=plastic)](../Benchmark/benchmark_methodology.md)
[![Benchmark](https://img.shields.io/badge/📊-Benchmark_Data-2a2a2a?style=plastic)](../Benchmark/symbolic_support_test.md)
[![Symbolic Prompting Class](https://img.shields.io/badge/Symbolic_Prompting-Class_2-blue?style=plastic&logo=youtube)](https://youtu.be/ey4t2T09L9I)


[⬅️ Class 1: What is Symbolic Prompting?](../BLOCK1_Fundamental/01_What_is_Symbolic_Prompting.md) | [🏠 Home](../README.md) | [Class 3: Normal vs. Atomic Tokens ➡️](../BLOCK1_Fundamental/03_Normal_vs_Atomic_Tokens.md)

</div>

---

<div align="center">
<center><img src="02_Prompt_as_State_Machine.png" alt="State machine diagram showing the transition between prompt states" style="width: 60%; height: auto;"></center>
</div>

## What happens BETWEEN one prompt and the next?

If you don’t control the state, you don’t control the system.

**What happens between one PROMPT and the next?** <br>
Artificial Intelligence does not live in a vacuum. Each prompt transforms something. That 'something' is called STATE.