A little about me
Hello, I am Qiming Li (启明 李), a master's student at HIT-SCIR, advised by Prof. Xiaocheng Feng and Prof. Bing Qin. I am currently a research intern with the Meituan Longcat Team. Previously, I was a visiting student at THUNLP, working with Prof. Zhiyuan Liu, and a research assistant at HKUNLP, working with Dr. Xiachong Feng and Prof. Lingpeng Kong. My research focuses on VLM mid- and post-training, including RL and On-Policy Distillation, Coding & GUI agents, and hallucination.
01 / TRAIN
VLM Post-Training
SFT & On-Policy Distillation
02 / ACT
Visual Agents
Coding & GUI Interaction
03 / TRUST
Reliable Perception
Hallucination & Interpretability
Where I've worked
Research Experience
Academic Research
- Mar.2025 - Sep.2025: HKUNLP, The University of HongKong, Research Intern
- Supervisor: Postdoc.Xiachong Feng and Prof.Lingpeng Kong
- Jul.2024 - Feb.2025: THUNLP, Tsinghua University, Research Intern
- Supervisor: Prof.Zhiyuan Liu
Industry Research
- May.2026 - now: Meituan, Longcat Team, Research Intern
- Jan.2026 - May.2026: Alibaba, Abot Team, Research Intern
- Oct.2025 - Jan.2026: Huawei, Xiaoyi, Research Intern
- July.2024 - Feb.2025: ModelBest, Research Intern
- Oct.2023 - Jan.2024: iFLYTEK, Research Intern
- Oct.2022 - Aug.2023: HITCRT, Vision Control Algorithm Engineer
- Jan.2022 - Feb.2022: Leju (Shenzhen) Robotics, Vision Algorithm Engineer
Research portfolio
Selected Publications
Google Scholar ↗* Equal contribution.
PRPO: Perception-Reinforced Policy Optimization via Token-Level Dynamic Advantage Reshaping
Not All Tokens See Equally: Perception-Grounded Policy Optimization for Large Vision-Language Models
CAST: Mitigating Object Hallucination in Large Vision-Language Models via Caption-Guided Visual Attention Steering
Causal Tracing of Object Representations in Large Vision Language Models: Mechanistic Interpretability and Hallucination Mitigation
CLAIM: Mitigating Multilingual Object Hallucination in Large Vision-Language Models with Cross-Lingual Attention Intervention
RLAIF-V: Open-Source AI Feedback Leads to Super GPT-4V Trustworthiness
Investigating and Mitigating the Multimodal Hallucination Snowballing in Large Vision-Language Models
Unlocking Multilingual Reasoning Capability of LLMs and LVLMs through Representation Engineering
MPR-GUI: Benchmarking and Enhancing Multilingual Perception and Reasoning in GUI Agents
Academic journey
Education
- B.S. in Computer Science, Harbin Institute of Technology, Aug.2020-June.2024
- M.S. in Computer Science, Harbin Institute of Technology, Aug.2024-2027(expected)
- Supervisor: Prof.Xiaocheng Feng
Giving back
Academic Service
- 2025: ACL ARR (Reviewer), AAAI (Program Committee), ICLR (Reviewer)
- 2026: ACL ARR (Reviewer), AAAI (Program Committee), ACM MM (Reviewer), NeurIPS (Reviewer), ICLR (Reviewer)
Teaching
- Teaching Assistant @ Harbin Institute of Technology
- 22CS31052 Knowledge Representation and Reasoning, Spring 2025
Milestones
Honors & Awards
- HIT “Sheng Sheng Bu Xi” Scholarship (No.1)
- HIT(ShenZhen) Insun Scholarship (Top-1%)
- Graduate Entrance Scholarship (Top Class)
- Graduate Academic Scholarship (Top Class)
- HIT Outstanding Graduate Student Award
- 2023 National Robotics Competition Robomaster
A global audience