1 citations · 1 across the 5 of their papers we have counts for
1 paper · 2 filters
Junze Zhou, David Klindt
Understanding when neural networks learn interpretable representations is central to mechanistic interpretability. We study this through identifiability: when can supervised learni…