Explore by interaction, not memorization.

See how a simple artificial neuron combines evidence to make a yes-or-no decision.

Follow signals through connected layers that learn increasingly useful patterns.

Watch a model parameter step downhill toward the lowest error.

Fit a line that turns one numerical input into a useful estimate.

Trace a series of plain-language questions that lead to a prediction.

Group unlabeled points by repeatedly moving cluster centers.

Predict from the examples that live nearest to a new data point.

Explore the architecture that lets modern language models compare every word in context.

See how a token decides which other tokens matter most right now.

Map meaning into coordinates so similar ideas sit near one another.