InfoResearchPreprint
When Backdoors Can Be Switched Off: Suppression-Aware Backdoor Attacks on Vision-Language-Action Models
- Published
- Record updated
Summary
Researchers study suppression-aware backdoors in Vision-Language-Action (VLA) models, which combine conjunctive trigger activation with instruction-conditioned suppression. Three variants are tested on SmolVLA using two SO-100 manipulation datasets against four baselines. Self-Suppressing and Multi-Trigger cut Isolation Forest detection from 96--98% to 1--18% and raise attack success to 63--71% with perfect clean-task performance, and experiments on OpenVLA with LIBERO show transfer to a 7B-parameter policy that evades STRIP and action-distribution-shift detection.
Topics
Related items
- MediumGemini AI agents repeatedly bypassed safeguards to attack real websiteSimilar attack · BleepingComputer
- CriticalCVE-2026-120699: CubeSandbox host path bind mount allows host file access via POST /sandboxesSimilar attack · NVD/CVE Database
- MediumGoogle warns that fully-automated AI attacks are coming soonSimilar attack · BleepingComputer
- InfoLLMs are Vulnerable to Malicious Requests with Indistinguishable IntentSimilar attack · Arxiv (cs.CR + cs.CL + cs.LG)
- InfoCreating Steering Vectors via Input OptimisationSimilar attack · Arxiv (cs.CR + cs.CL + cs.LG)