About

I am seeking opportunities: I will graduate in June 2027, and am looking for research or engineering roles in LLMs, post-training, safety alignment, and agents. Reach me at 15738183237@163.com, or WeChat Ssyllzz.

I am currently pursuing my master's degree in Computer Science and Technology at the Institute of Software, CAS, with an expected graduation in June 2027. Before that, I received my B.Eng. in Computer Science and Technology from Zhengzhou University in 2024.

News

  • Apr 2026 Our paper "When Models Outthink Their Safety: Unveiling and Mitigating Self-Jailbreak in Large Reasoning Models" has been accepted to ACL 2026 Findings!
  • Apr 2026 Received CANN (Huawei Ascend) Core Contributor certification.
  • May 2025 Contributed to AutoAlign project, an open-source toolkit for automated LLM alignment.

Research Interests

  • Large Language Models
  • Post-training
  • Safety Alignment
  • Agents

Education

2020.09 - 2024.06

Zhengzhou University

B.Eng. in Computer Science and Technology

Internships

Jun 2026 - Present

Ant Group · Alipay Business Group, Edge Intelligence Department

Algorithm Intern

Participated in the development and iterative optimization of AI Alipay (Abao), helping analyze and resolve model issues in real-world user scenarios.

Selected Publications

When models outthink their safety: Mitigating self-jailbreak in large reasoning models with chain-of-guardrails

Yingzhi Mao, Chunkang Zhang, Junxiang Wang, Xinyan Guan, Boxi Cao, Yaojie Lu, Hongyu Lin, Xianpei Han, Le Sun

ACL 2026 Findings · Paper · Code

AUTOALIGN: Get Your LLM Aligned with Minimal Annotations

ACL Demo 2025 · Paper · Code

Honors

  • 2026 CANN (Huawei Ascend) Core Contributor Certification
  • 2020 Zhengzhou University Outstanding Student Scholarship

WeChat

Scan the QR code to add me on WeChat.

WeChat QR code for Yingzhi Mao

WeChat ID: Ssyllzz