An open–source anti–deskilling framework built on a decade of human–AI collaboration research from the Cabitza lab.

Emotions live as geometric directions inside a model’s hidden states.

A model that looks well–calibrated globally can be broken exactly where it matters.