alphaFeedback
MZ

Mohammad Zandsalimy

cs.CRcs.AIcs.LGcs.CL

On Valency

published · living versions
W_8bxezcmd·v1 · currentpublished

Preprints & journals

9 papers in the corpus · 2025–2026
Refusing Everything Looks Safe: Restoring the Benign Arm to Encoded-Prompt Evaluation2609.26176v2 · Haoyu Zhang, Haowen Xu, Xiao Luo et al.2026 · 0 citationsarXiv
Unread or Unenforced? Separating Representation from Enforcement Failure in Content Guards2609.26178v3 · Haoyu Zhang, Yi Feng, Shibo Zheng et al.2026 · 0 citationsarXiv
0%, 45%, or 99%: A Guardrail's Own Share of the Refusals It Is Credited With2608.08641v2 · Haoyu Zhang, Xiangchen Guan, Yang Chen et al.2026 · 0 citationsarXiv
Decoy Images Amplify Caption-Mediated Defenses Against Encoded Jailbreaks2608.01043v2 · Haoyu Zhang, Xiangchen Guan, Shibo Zheng et al.2026 · 0 citationsarXiv
Depth, Not Breadth: Best-of-N Jailbreaking Beyond Surface Noise2607.26639v3 · Haoyu Zhang, Hanwen Liu, Yang Chen et al.2026 · 0 citationsarXiv
Exposing LLM Safety Gaps Through Mathematical Encoding:New Attacks and Systematic Analysis2605.03441v1 · Haoyu Zhang, Mohammad Zandsalimy, Shanu Sushmita2026 · 0 citationsarXiv
Behind the Mask: Benchmarking Camouflaged Jailbreaks in Large Language Models2509.05471v1 · Youjia Zheng, Mohammad Zandsalimy, Shanu Sushmita2025 · 0 citationsarXiv
Career total: 26 works. 9 are in this corpus.

Profile built from the corpus for this byline.

Author records are still filling in while the Hub is in alpha. If this is your page, you'll be able to claim it soon. Spot a mistake? Tell us.