Qingyu Yin

↑ Click my name for 中文

Research

Currently focused on scaling RL and alignment.

Education

Zhejiang University
2022.09 - 2026.06
B.A. in English (Linguistics)

Experience

StepFun · Post-Training Team
2026 - Present
MiniMax · Post-Training Team TOP Talent Intern
2026.03 - 2026.08
RL for M3 and M3.1.
JD Group, JoyAI · Post-Training Team JD Young Intern
2025.11 - 2026.02
Built the RL pipeline for JoyAI-LLM-Flash from scratch as one of three core members.
Rednote Hi Lab · Post-Training Team Ace Top Talent Intern
2025.07 - 2025.11
Post-training of dots.llm2 / dots.vlm2.

Publications

10 papers · 7 first / co-first author · 2 oral presentations

Evaluating Parameter Efficient Methods for RLVR
ICML 2026
Qingyu Yin*, Yulun Wu*, Zhennan Shen*, Sunbowen Li, Zhilin Wang, Yanshu Li, Chak Tou Leong, Jiale Kang, Jinjin Gu.
Refusal Falls off a Cliff: How Safety Alignment Fails in Reasoning?
COLM 2026
Qingyu Yin, Chak Tou Leong, Linyi Yang, Wenxuan Huang, Wenjie Li, Xiting Wang, Jaehong Yoon, Jinjin Gu.
MiSS: Balancing LoRA Performance and Efficiency with Simple Shard Sharing
ICLR 2026
Jiale Kang*, Qingyu Yin*.
Constrain Alignment with Sparse AutoEncoders
ICML 2025
Qingyu Yin, Chak Tou Leong, Hongbo Zhang, Minjun Zhu, Hanqi Yan, Qiang Zhang, Yulan He, Wenjie Li, Jun Wang, Yue Zhang.
ModRWKV: Transformer Multimodality in Linear Time
EMNLP 2025 🏆 Oral
Qingyu Yin*, Jiale Kang*, Ziyin Yue*, Jiang Rui, Weile Li, Zening Lu, Zhouran Ji.
StableMask: Refining Causal Masking in Decoder-only Transformers
ICML 2024
Qingyu Yin, Xuzheng He, Xiang Zhuang, Yu Zhao, Jianhua Yao, Xiaoyu Shen, Qiang Zhang.
Deeper Insights Without Updates: The Power of In-Context Learning Over Fine-Tuning
EMNLP 2024 Findings
Qingyu Yin, Xuzheng He, Luoao Deng, Chak Tou Leong, Fan Wang, Yanzhao Yan, Xiaoyu Shen, Qiang Zhang.
Vision-DeepResearch: Incentivizing DeepResearch Capability in Multimodal Large Language Models
ICML 2026
Wenxuan Huang, Yu Zeng, Qiuchen Wang, Zhen Fang, Shaosheng Cao, Zheng Chu, Qingyu Yin, Shuang Chen, Zhenfei Yin, Lin Chen, Zehui Chen, Xu Tang, Yao Hu, Philip Torr, Feng Zhao, Wanli Ouyang.
Interleaving Reasoning for Better Text-to-Image Generation
ICLR 2026
Wenxuan Huang, Shuang Chen, Zheyong Xie, Shaosheng Cao, Shixiang Tang, Yufan Shen, Qingyu Yin, Wenbo Hu, Xiaoman Wang, Yuntian Tang, et al.
Aligned LLMs’ Safety Mechanisms Tend to Be Anchored in The Template Region
ACL 2025 🏆 Oral
Chak Tou Leong, Qingyu Yin, Jian Wang, Wenjie Li.

* Equal contribution.