Kexun Zhang

I am an AI researcher currently interested in LLM for code generation and coding agents.

I used to be a competitive programmer and an amateur Xiangsheng performer for fun.

Email / GitHub / Scholar / Twitter

Selected Research

* denotes equal contribution. † denotes project lead.

	Is Vibe Coding Safe? Benchmarking Vulnerability of Agent-Generated Code in Real-World Tasks 微博编码安全不？测量脆弱性属于经理人生成代码在真实任务中 Songwen Zhao, Danqing Wang†, Kexun Zhang†, Jiaxuan Luo, Zhuo Li, Lei Li preprint, 2025 paper / code /
	HardTests: Synthesizing High-Quality Test Cases for LLM Coding 硬测：合成高质量测试盒子为了大语言模型写代码 Zhongmou He, Yee Man Choi, *Kexun Zhang†*, et al. preprint*, 2025 paper / code / website /
	Integrating Expertise of Software Engineering Agents 整合专长来自软件工程特工 Kexun Zhang, Weiran Yao, Zuxin Liu, et al. ICLR, 2025 paper / code /
	Scaling LLM Inference with Optimized Sample Compute Allocation 扩大大语言模型推理使用优化过的样本计算分配 Kexun Zhang, Shang Zhou, Danqing Wang, William Yang Wang, Lei Li NAACL, 2025 paper / code /
	SWE-Search: Enhancing Software Agents with Monte Carlo Tree Search and Iterative Refinement 穗搜：加强软件特工使用莫奈特卡洛树搜索和逐步改进 Antonis Antoniades, Albert Örwall, Kexun Zhang, Yuxi Xie, Anirudh Goyal, William Wang ICLR, 2025 paper / code /
	Invisible Image Watermarks Are Provably Removable Using Generative AI 隐形图片水印能被人工智能可证明地移除 Xuandong Zhao, Kexun Zhang*, Zihao Su, Saastha Vasan, Ilya Grishchenko, Christopher Kruegel, Giovanni Vigna, Yu-Xiang Wang, Lei Li NeurIPS*, 2024 paper / code /
	Hire a Linguist!: Learning Endangered Languages with In-Context Linguistic Descriptions 雇佣一个语言学家！学习濒危语言通过上下文语言描述 Kexun Zhang, Yee Man Choi, Zhenqiao Song, Taiqi He, William Yang Wang, Lei Li Findings of ACL, 2024 paper / code /
	ToolDec: Syntax Error-Free and Generalizable Tool Use for LLMs via Finite-State Decoding 兔耳德克：语法错误没有的而且通用的大语言模型工具使用通过有限状态解码 Kexun Zhang, Hongqiao Chen, Lei Li, William Yang Wang NeurIPS MathAI Workshop, 2023 paper / code /
	ALGO: Synthesizing Algorithmic Programs with LLM-Generated Oracle Verifiers 阿尔戈：合成算法程序使用大语言模型生成的神谕验证器 Kexun Zhang, Danqing Wang, Jingtao Xia, William Yang Wang, Lei Li NeurIPS, 2023 paper / code /
	ReDi: Trajectory Retrieval for Learning-Free Efficient Diffusion Inference 热滴：无学习轨迹检索用于高效扩散推断 Kexun Zhang, Xianjun Yang, William Yang Wang, Lei Li ICML, 2023 paper / code /
	A Study of Syntactic Multi-Modality in Non-Autoregressive Machine Translation Kexun Zhang, Rui Wang, Xu Tan, Junliang Guo, Yi Ren, Tao Qin, Tie-Yan Liu NAACL, 2022 paper /
	WSRGlow: A Glow-based Waveform Generative Model for Audio Super-Resolution Kexun Zhang, Yi Ren, Changliang Xu, Zhou Zhao Interspeech, 2021 paper / code / audio samples / Employed generative flow model on the speech super-resolution task, achieved state-of-the-art performance. Nominated for the Best Student Paper Award.

Design and source code from Jon Barron's website