Kuyun AI · Works to try

LLM Model and Training: From Tokens to Aligned Assistants

From the token up: twelve lessons, a midterm and a final, each with worked explanations and practice.

course · by @tech-courses

How large language models are built, trained, aligned, run, and used well — from the transformer and pretraining to fine-tuning, alignment, inference, and responsible use.

What's inside · 14

  1. Lesson 1 — What a Language Model Is
  2. Lesson 2 — The Transformer and Attention
  3. Lesson 3 — Inside the Model: Embeddings, Positions, Heads
  4. Lesson 4 — Pretraining and Scaling Laws
  5. Lesson 5 — Compute, Memory, and Parallelism
  6. Lesson 6 — Fine-Tuning and Adaptation
  7. Lesson 7 — Midterm: Model and Training
  8. Lesson 8 — Alignment: Making Models Helpful and Safe
  9. Lesson 9 — Inference: Getting Answers Out Efficiently
  10. Lesson 10 — Evaluating Language Models
  11. Lesson 11 — Using LLMs: Prompting, Context, and Retrieval
  12. Lesson 12 — Reasoning and Test-Time Compute
  13. Lesson 13 — Risks, Limits, and Responsible Use
  14. Lesson 14 — Final: Model and Training, End to End

Sign in, or sign up if you are new — either way you land on this work.