Attribution robustness arises implicitly from SGD dynamics through parameter-input curvature links but fails for softmax attention due to entropy constraints and is restored by kernel-based attention.
Title resolution pending
1 Pith paper cite this work, alongside 1 external citations. Polarity classification is still indexing.
1
Pith paper citing it
1
external citations · OpenAlex
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Improving Adversarial Robustness of Attribution via Implicit Regularization
Attribution robustness arises implicitly from SGD dynamics through parameter-input curvature links but fails for softmax attention due to entropy constraints and is restored by kernel-based attention.