Blog

Building physical intelligence through interpretability

How understanding what models learn can improve physical generalisation and enable more capable, robust and safer foundation models.

Read article