Visual Labs

Small interactive instruments for understanding AI systems.

These demos are intentionally compact: each one turns a core idea into something you can inspect, adjust, and reason about.

Model Mechanics

Explore how logits become a probability distribution.

Optimization

Visualize how optimization moves across an objective surface.

Representation

Inspect how text is broken into tokens and embedded for models.

Transformers

See how transformer attention links tokens through query-key-value structure.

Reliability

Explore model behavior outside the shape of training data.

Step through how search trees are expanded, simulated, and scored.

Evaluation

Compare loss behavior as predicted probabilities change.