Interactive visual lab · LLM fundamentals

Inside the Answer: Follow the Signal

Watch a question travel through an LLM pipeline—from raw text to generated answer. Select any stage to inspect its job, inputs, outputs, and live example.

Example input
“Why do cats purr?”
Illustrative flow · simplified for learning
Ready · Stage 1 of 8

Eight-stage question-to-answer pipeline

Click a stage · follow the glowing signal →

questiontoken IDsvectors
01

Input text

The journey begins with a human-readable question.

Input: Output: Stage: 1 / 8

Watch the representation change

A conceptual view of the data at the selected stage.

Active signalRepresentationOutput / next stage
Go deeper into LLMs

You have just followed a question through a full LLM pipeline. Each subtopic below looks at one thing that makes a model large and capable: its size, architecture, data, compute and inference.

8 subtopics — click to study
What Makes a Model Large? 🔴 Parameters 🔴 Model Architecture 🔴 Context Length 🔴 Training Data 🟡 Training Compute 🟡 Scaling 🟡 Inference 🔴
← Back to Roadmap
This page 0:00
This session 0:00
Total (all pages) 0:00
Today 0:00
▶ Running
📊 View all study stats →
0:00 studying