PRAC redirects Computer Use Agents' product selection by concentrating attention on an adversarial patch in the vision input, generalizing from white-box creation to fine-tuned models.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
verdicts
UNVERDICTED 2representative citing papers
GTA1 combines test-time scaling for action plan selection with RL-based grounding to achieve SOTA results on GUI agent benchmarks.
citing papers explorer
-
Preference Redirection via Attention Concentration: An Attack on Computer Use Agents
PRAC redirects Computer Use Agents' product selection by concentrating attention on an adversarial patch in the vision input, generalizing from white-box creation to fine-tuned models.
-
GTA1: GUI Test-time Scaling Agent
GTA1 combines test-time scaling for action plan selection with RL-based grounding to achieve SOTA results on GUI agent benchmarks.