Introduction to LLM Agents
Chapter 1: What Are LLM Agents?
From LLMs to Intelligent Action
Moving Past Simple Chatbots
What Exactly Is an LLM Agent?
The Purpose of LLM Agents
Comparing Agents to Scripts
Illustrative Agent Applications
Understanding Agent Independence
Chapter 2: The Building Blocks of an LLM Agent
The Central Role of the LLM
Crafting Instructions for Agents
Giving Agents Abilities
Remembering Recent Events
Thinking Before Acting
The Agent's Operational Loop
A Simplified Agent Workflow
Chapter 3: Your First LLM Agent: A Step-by-Step Guide
Preparing Your Workspace for Agent Development
Selecting an LLM for Your First Agent
Instructing Your Agent for a Task
Specifying the Agent's Goal
Coding the Agent's First Action
Executing and Monitoring Your Agent
Addressing Initial Problems
Hands-on Practical: Creating a To-Do List Agent
Chapter 4: Equipping Agents With Tools
Understanding Agent Tools
The Necessity of Tools for Agents
Creating a Basic Tool: A Calculator Example
Connecting Tools to Your Agent
Prompting Agents to Utilize Tools
Agent Tool Selection Logic
Survey of Available Tool Categories
Hands-on Practical: Equipping Your Agent with Search
Chapter 5: Basic Agent Planning and Execution
What is Agent Planning?
Guiding Agent Reasoning with Chain-of-Thought
Introducing the ReAct Approach
Setting Clear Objectives for Agents
Decomposing Complex Tasks
Tracking Task Execution
Managing Basic Execution Failures
Practice: Building a ReAct Agent for Sequential Tasks
Chapter 6: LLM Agent Memory: Remembering Information
Why Agents Need to Remember
A Look at Different Memory Systems
Focusing on Conversational Context
A Simple Implementation of Short-Term Memory
The Influence of Memory on Agent Behavior
Understanding Short-Term Memory Boundaries
Hands-on Practical: Adding Contextual Memory to Your Agent
Chapter 2: The Building Blocks of an LLM Agent
After establishing what LLM agents are, we now look at their internal structure. This chapter breaks down an LLM agent into its essential components. You will learn about:
- The Large Language Model (LLM) serving as the agent's cognitive core.
- How instructions and prompts are crafted to direct agent behavior.
- The function of tools in extending an agent's capabilities to interact with external systems.
- The concept of short-term memory for maintaining context during operations.
- Basic methods an agent uses for planning and sequencing actions.
- The cyclical process of observation, thought, and action that defines an agent's operation.
Understanding these individual parts and how they fit together is key to grasping how agents process information and accomplish goals.
Sections
2.1 The Central Role of the LLM
2.2 Crafting Instructions for Agents
2.3 Giving Agents Abilities
2.4 Remembering Recent Events
2.5 Thinking Before Acting
2.6 The Agent's Operational Loop
2.7 A Simplified Agent Workflow
© 2025 ApX Machine Learning