Definition
This page connects the main generative AI terms. An LLM generates text. A prompt is what you ask it. RAG lets it answer from your documents. An agent lets it use tools and act. Fine-tuning adapts a model with your own examples.
In simple words
The LLM is a smart employee. The prompt is your instruction. RAG hands them the company files. Agents give them a computer and permission to do things. Fine-tuning is job training.
Why it matters
These words get thrown around constantly. With this map, you can place each one, and go to its full module when you need to.
How it works
| Term | Answers | Use it when |
|---|---|---|
| LLM | What generates the text? | Always, as the base |
| Prompt | How do I ask? | Always. Try this first. |
| RAG | How does it know my data? | Answers must come from your documents |
| Agent | How does it take actions? | Steps can't be fixed in advance |
| Fine-tuning | How does it learn my style? | Prompts and RAG aren't enough |
| Embedding | How is meaning stored? | Search and RAG |
Common mistakes
- Fine-tuning when you really needed RAG. To add facts, use RAG.
- Building an agent for a task with fixed steps.
Check yourself
Summary
- LLM generates, a prompt asks, RAG adds knowledge, and agents act.
- Start with the simplest thing that works.
Finished reading?