Built with and by Teycir Ben Soltane•
How to Use•FAQ•GitHub•arXiv.org•
Share:
ArXivCSExplorer
☆☆Bookmarks🏆RSSHow to UseFAQ
Home/Authors/Dandan Tu

Dandan Tu

2 indexed papers

Recent (6 mo)
2
With code
0
Influential cites
0
Benchmarked
0

Publications per year

2
26

Top categories

AI×2Vision×1Crypto×1

Frequent co-authors

Ya-Qi Yu1×
Hao Wang1×
Fangyu Hong1×
Xiangyang Qu1×
Gaojie Wu1×
Qiaoyu Luo1×

Research Timeline

2026
Safety Geometry Collapse in Multimodal LLMs and Adaptive Drift Correction

This paper introduces the concept of Safety Geometry Collapse, demonstrating that multimodal inputs degrade the safety separation of LLMs, and proposes ReGap, a training-free method that adaptively corrects this drift to improve MLLM safety.

Reinforcement Learning with Robust Rubric Rewards

The paper introduces $ ext{RLR}^3$, a novel framework that extends verifiable rewards in Reinforcement Learning to handle partially verifiable, multi-criteria vision-language tasks by integrating robust rubric scoring.

Highlighted terms show continued research focus across papers

Papers

cs.CVcs.AIRecentMay 28, 2026

Reinforcement Learning with Robust Rubric Rewards

Ya-Qi Yu, Hao Wang, Fangyu Hong, Xiangyang Qu +14 more

The paper introduces $ ext{RLR}^3$, a novel framework that extends verifiable rewards in Reinforcement Learning to handle partially verifiable, multi-criteria vision-language tasks by integrating robu…

View →
cs.AIcs.CRRecentMay 18, 2026

Safety Geometry Collapse in Multimodal LLMs and Adaptive Drift Correction

Jiahe Guo, Xiangran Guo, Jiaxuan Chen, Weixiang Zhao +5 more

This paper introduces the concept of Safety Geometry Collapse, demonstrating that multimodal inputs degrade the safety separation of LLMs, and proposes ReGap, a training-free method that adaptively co…

View →