OpenSAE

non-profit

Building open foundations for understanding, evaluating, and intervening in neural networks.

Hunter26666666 updated a collection about 3 hours ago

Hunter26666666 updated a collection about 3 hours ago

Hunter26666666 updated a Space about 3 hours ago

Behavioral Compositionality of Sparse Autoencoder Representations

OpenSAE 's Spaces 3

Explore verified SAE checkpoints, interventions and results.

Open infrastructure for mechanistic interpretability.

Five-minute tutorial on SAE behavioral compositionality.