ResearchTrend.AI
  • Papers
  • Communities
  • Organizations
  • Events
  • Blog
  • Pricing
  • Feedback
  • Contact Sales
Papers
Communities
Social Events
Terms and Conditions
Pricing
Contact Sales
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2508.17456
24
0
v1v2 (latest)

Adversarial Examples Are Not Bugs, They Are Superposition

24 August 2025
Liv Gorton
Owen Lewis
    AAML
ArXiv (abs)PDFHTML
Main:9 Pages
9 Figures
Bibliography:4 Pages
2 Tables
Appendix:1 Pages
Abstract

Adversarial examples -- inputs with imperceptible perturbations that fool neural networks -- remain one of deep learning's most perplexing phenomena despite nearly a decade of research. While numerous defenses and explanations have been proposed, there is no consensus on the fundamental mechanism. One underexplored hypothesis is that superposition, a concept from mechanistic interpretability, may be a major contributing factor, or even the primary cause. We present four lines of evidence in support of this hypothesis, greatly extending prior arguments by Elhage et al. (2022): (1) superposition can theoretically explain a range of adversarial phenomena, (2) in toy models, intervening on superposition controls robustness, (3) in toy models, intervening on robustness (via adversarial training) controls superposition, and (4) in ResNet18, intervening on robustness (via adversarial training) controls superposition.

View on arXiv
Comments on this paper