The Playgrounds.
Less theory. More “what happens if?”
Pull things apart. Move the numbers.
Get a feel for how AI actually works.
A little less magic.
A lot more understanding.
Choose a sentence and follow its words through a toy language model, from numeric representations to attention and prediction.
Open the playground How LLMs Work: The Full PipelineFollow your curiosity.
Pick an experiment. Change something. See what clicks.
Understanding Backpropagation
Follow one mistake back through a network.
Explore playgroundThe Blind Hiker, Gradient Descent Playground
Find out how learning goes wrong.
Explore playgroundBackpropagation
Follow one mistake back through a network.
Explore playgroundCNN Builder
Build something. Then look inside it.
Explore playgroundCNN Layer Explorer
See the same image nine different ways.
Explore playgroundNeural Network Digit Classifier
Can you make a digit ambiguous?
Explore playgroundLinear Regression
Can you beat your first fit?
Explore playgroundLogistic Regression
Find the edge between two classes.
Explore playgroundDecision Tree
Find every route through the tree.
Explore playgroundRandom Forest
Get the experts to disagree.
Explore playgroundK-Means Clustering
Watch order emerge from a crowd.
Explore playgroundOverfitting
Win on data you have never seen.
Explore playgroundNeural Network
Make three simple voices draw a boundary.
Explore playgroundRecurrent Neural Network
See how long a word survives.
Explore playgroundTransformer
Follow the connections between words.
Explore playgroundGAN
Turn noise into a recognizable pattern.
Explore playgroundReinforcement Learning
Find a route, then teach the robot.
Explore playground