LLMs Deep Dive
22 chapters · LLM Science

A field guide toLarge Language Models

How large language models work — architecture, pretraining, alignment, and the frontier.

chapters
22chapters
practice questions
980practice questions
coding problems
69coding problems
What's inside

Reading, until it actually sticks

01

AI code review

Submit a solution, get a focused critique — correctness, style, complexity — in seconds.

Submissionsoftmax.py
1def softmax(x):
2 e = np.exp(x - x.max())
3 return e / e.sum()
All 5 tests passed
02

On-demand AI tutor

Stuck on a paragraph? One click and a tutor re-explains it with more intuition, more math, or a fresh angle.

Tutorshort

What is attention?

Attention weighs tokens by relevance.

03

Progress tracking

Chapters read, questions understood, problems solved — quietly tracked as you go.

Chapter 013/4
  • TokenizationRead
  • AttentionRead
  • Positional encodingsRead
  • Layer normNext
The syllabus

Everything, in six parts

Part I

Foundations

4 chapters
Part II

Pretraining & Scale

3 chapters
Part III

Architecture Variants

3 chapters
Part IV

Post-Training, Alignment & Evaluation

6 chapters
Part V

Applications

4 chapters
Part VI

Frontiers

2 chapters
Also on MLGym