A dataless regularizer based on Kronecker-Factored Approximate Curvature disentangles task vectors for improved task addition and negation in foundation models, with constant complexity and robustness to rescaling.
This property is further enhanced under our regularization regime, where only a few darker regions remain, mostly forα >1, a setting that is never used in practice
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Dataless Weight Disentanglement in Task Arithmetic via Kronecker-Factored Approximate Curvature
A dataless regularizer based on Kronecker-Factored Approximate Curvature disentangles task vectors for improved task addition and negation in foundation models, with constant complexity and robustness to rescaling.