Mingye Zhu (祝铭烨)

Ph.D. Candidate (Master-to-Ph.D. transfer) in College of Cyberspace Security, University of Science and Technology of China (USTC), Hefei, China.

Research focus: LLM alignment.

Advisors: Prof. Yongdong Zhang, Prof. Zhendong Mao.

Email: steviezhu1 at gmail.com

About

I am a Ph.D. candidate at USTC focusing on alignment, safety, and trustworthy behavior of large language models. My work spans alignment training, robust optimization, and inference-time alignment.

I am currently on the job market.

Research Interests

Research experience.

Selected Publications

  1. Zhu, M., Liu, Y., Fu, Z., Wang, Q., & Mao, Z. (2025). In-Token Rationality Optimization: Towards Accurate and Concise LLM Reasoning via Self-Feedback. AAAI 2026 (Oral).

  2. Zhu, M., Liu, Y., Fu, Z., Zhang, Y., & Mao, Z. (2025). Leveraging Robust Optimization for LLM Alignment under Distribution Shifts. NeurIPS 2025.

  3. Zhu, M., Liu, Y., Zhang, L., Guo, J., & Mao, Z. On-the-fly Preference Alignment via Principle-Guided Decoding. ICLR 2025.

Full list of publications.

Education

Awards and Activities

Honors, internships, and activities.