Developmental interpretability

The study of how interpretable representations, mechanisms, and circuits form and change during training.

The study of how interpretable representations, mechanisms, and circuits form and change during training.

is aResearch area All entries Open in graph

It connects explanations of a trained model with the trajectory through which its internal structure emerged.

Connections

subarea of

Mechanistic interpretability Training dynamics

Related publications

addressed by

Incremental Learning of Sparse Attention Patterns in Transformers

2026