Stanford CME 295 Independent interactive companion
CME 295 Transformers &
Large Language Models
Visit the course website
CME 295 / From theory to practice

Interactive illustrations

Change a parameter. Follow the computation. Explore attention, sampling and masked diffusion here, separately from the chapter reading and quizzes.

Three teaching experiments

Explore the mechanics

Toy inputs. Real calculations. Visible steps.

Attention and temperature use invented vectors and fixed logits; masked diffusion follows a scripted schedule. No trained models or AI API calls are involved.

Back to the cheatsheet and chapter quizzes