Interpreting Language Model Parameters
The method, adVersarial Parameter Decomposition (VPD), optimizes for decompositions of neural network parameters into simple subcomponents that preserve the network's input-output behavior even when many subcomponents are ablated, including under ablations that are adversarially selected to destroy behavior.
Lucius Bushnaq, Dan Braun, Oliver Clive-Griffin et al.
· 4 citations
· ⚡1