This visual primer builds the mathematical intuition needed to follow how language models learn and how a transformer processes information. It is meant as a practical refresher rather than a formal math course.
Inside the reader
- Foundations 5 chapters
- Vectors, matrices, probability, loss, gradients, and the role each plays in model training.
- Transformers 1 chapter
- A guided pass through attention, MLPs, residual connections, and normalization.
- Reading papers 1 chapter + references
- Common mathematical moves, a Greek-letter guide, glossary, and curated learning resources.