SecureWebArena is a new benchmark suite for holistic security evaluation of LVLM-based web agents using diverse simulated environments, attack taxonomies, and multi-layered failure analysis across reasoning, behavior, and outcomes.
arXiv preprint arXiv:2503.15092 (2025)
4 Pith papers cite this work. Polarity classification is still indexing.
verdicts
UNVERDICTED 4representative citing papers
Training large reasoning models only on safety verification tasks internalizes safety understanding and boosts robustness to out-of-domain jailbreaks, providing a stronger base for reinforcement learning alignment than standard supervised fine-tuning.
PRJA achieves 83.6% average success injecting harmful content into LRM reasoning chains on five QA datasets without altering final answers.
PRISM decomposes harmful instructions into benign visual gadgets and directs LVLMs via prompts to compose them through reasoning into harmful outputs, achieving ASR over 0.90 on SafeBench.
citing papers explorer
-
SecureWebArena: A Holistic Security Evaluation Benchmark for LVLM-based Web Agents
SecureWebArena is a new benchmark suite for holistic security evaluation of LVLM-based web agents using diverse simulated environments, attack taxonomies, and multi-layered failure analysis across reasoning, behavior, and outcomes.
-
Internalizing Safety Understanding in Large Reasoning Models via Verification
Training large reasoning models only on safety verification tasks internalizes safety understanding and boosts robustness to out-of-domain jailbreaks, providing a stronger base for reinforcement learning alignment than standard supervised fine-tuning.
-
Reasoning-targeted Jailbreak Attacks on Large Reasoning Models via Semantic Triggers and Psychological Framing
PRJA achieves 83.6% average success injecting harmful content into LRM reasoning chains on five QA datasets without altering final answers.
-
PRISM: Programmatic Reasoning with Image Sequence Manipulation for LVLM Jailbreaking
PRISM decomposes harmful instructions into benign visual gadgets and directs LVLMs via prompts to compose them through reasoning into harmful outputs, achieving ASR over 0.90 on SafeBench.