Hi there Welcome to my Homepage!

I am a CS Ph.D. student in the ASPIRE Lab at ShanghaiTech University, advised by Prof. Wenjie Wang. Prior to that, I received my bachelorโ€™s degree from Northwest A&F University in 2024.

My research focuses on AI safety, alignment, and post-training for LLMs and VLMs. Recently, I have been exploring latent space reasoning and looped language models (LoopLMs), with an emphasis on efficient post-training and safety alignment.

News

  •  ๐ŸŽ‰๐ŸŽ‰ SaLR was accepted to NeurIPS 2026!
  •  ๐ŸŽ‰๐ŸŽ‰ One paper was accepted to ACML 2026!
  •  ๐ŸŽ‰๐ŸŽ‰ One paper was accepted to NLPCC 2026!
  •  ๐ŸŽ“๐ŸŽ“ I officially became a Ph.D. student at ShanghaiTech University!
  •  ๐ŸŽ‰๐ŸŽ‰ EVA was accepted for publication in IEEE TPAMI 2026!
  •  ๐ŸŽ‰๐ŸŽ‰ DELMAN was accepted to ACL 2025!
  •  ๐ŸŽ“๐ŸŽ“ I began my master's studies at ShanghaiTech University!

Publications

Preprint Overview of dynamic cross-loop on-policy distillation for looped language models

Recurrent Self-Improvement: Dynamic Cross-Loop On-Policy Distillation for Looped Language Models

Yi Wang, Rui Qian, Yu Li, Haoyang Yao, Wenjie Wang†.

Preprint

Preprint Safety challenges across recurrent depths in looped language models

Safe at One Loop, Risky at Another: Aligning Safety Across Recurrent Depths in Looped Language Models

Yi Wang*, Xiuyuan Qi*, Dongqi Han, Dongsheng Li, Wenjie Wang†.

Preprint

NeurIPS 2026 Overview of Safety-aware Latent Space Reasoning (SaLR)

Safety-Aware Latent Space Reasoning in Large Language Models

Yi Wang*, Wenjie Wang*†, Hongye Qiu, Yu Pan.

NeurIPS 2026

NLPCC 2026 Overview of DR.GAP: Gender-aware prompting with decoupled reasoning

DR.GAP: Mitigating Bias in Large Language Models using Gender-Aware Prompting with Decoupled Reasoning

Hongye Qiu*, Yue Xu*, Yi Wang, Meikang Qiu, Wenjie Wang†.

NLPCC 2026

ACML 2026 Differential privacy training workflow for preference alignment on survey-derived data

Differential Privacy Protected Preference Alignment on Survey Derived Data

Xiuyuan Qi, Yi Wang, Wenjie Wang†.

ACML 2026

Preprint Response patterns of diffusion and autoregressive vision-language models in GPO-V

GPO-V: Jailbreak Diffusion Vision Language Model by Global Probability Optimization

Yu Pan, Andi Zhang, Yi Wang, Sibei Yang, Wenjie Wang†.

Preprint

ACL 2025 Overview of DELMAN: Dynamic defense against language model jailbreaking with model editing

DELMAN: Dynamic Defense Against Large Language Model Jailbreaking with Model Editing

Yi Wang, Fenghua Weng, Sibei Yang, Zhan Qin, Minlie Huang, Wenjie Wang†.

ACL 2025

Education

2024.09 - Present
ShanghaiTech University
Ph.D. Student
2020.09 - 2024.06
Northwest A&F University
Undergraduate

Experience

2026.06 - 2026.09
Microsoft Research Asia (MSRA)
Research Collaboration

Services

  • [Conference Reviewer]
    • ICLR 2027
    • ACL ARR 2025, 2026
  • [Journal Reviewer]
    • Information Fusion๏ผˆไธญ็ง‘้™ขไธ€ๅŒบ TOP๏ผŒJCR Q1๏ผ‰
    • Pattern Recognition๏ผˆไธญ็ง‘้™ขไธ€ๅŒบ TOP๏ผŒJCR Q1๏ผ‰
  • [Open Source Contributor]

Awards

  • Outstanding Student
  • ๐Ÿ…Provincial First Prize, 15th National College Student Mathematics Competition
  • Academic Scholarship
  • Outstanding Student Leader