Cunxiang Wang PhD ZhipuAI & Tsinghua University Email: wangcunxiang303 AT gmail.com, wangcunxiang AT foxmail.com

Biography

I am a tech leader at ZhipuAI and a postdoc at Tsinghua University, fortunately working with Prof. Jie Tang and Prof. Minlie Huang. Previously, I obtained my Ph.D. at Zhejiang University & Westlake University joint Ph.D. program, and was fortunately supervised by Prof. Yue Zhang, WestlakeNLP Lab.

My current research primarily focuses on Large Language Models, especially LLM Evaluation and General Agent.

We are looking for self-motivated interns and full-time employees to work with us at Tsinghua or ZhipuAI. We will offer cutting-edge technology and insights for LLMs, adequate computing and data resources, a competitive salary, and a research-friendly environment. If you are interested, please feel free to contact me.

News

Education

• Westlake University & Zhejiang University Co-PhD Promgram
  Aug, 2019 - Jun, 2024 (Expected)
  
  Ph.D candidate
  Supervisor: Yue Zhang
  Degree Granted By: Zhejiang University
  
• Dept. of Computer Science and Technology, Northeastern University, Shenyang
  Sep, 2015 - Jul, 2019
  
  Undergraduate student.
  Interned at Northeastern NLP Lab, 2016-2019, advised by Prof. Feiliang Ren
  

Experience

• Amazon AWS AI
  March, 2022 - now
  
  Research Intern
  Mentor: Qipeng Guo and Zheng Zhang
  

Selected Publications (As First / Co-First Author) [Google Scholar] [Semantic Scholar]

* indicates co-first author.
† indicates corresponding author.

2025

• NovelQA: Benchmarking Question Answering on Documents Exceeding 200K Tokens [pdf][Github Page]
  ICLR-2025
  Cunxiang Wang*, Ruoxi Ning*, Boqi Pan, Tonghui Wu, Qipeng Guo, Cheng Deng, Guangsheng Bao, Qian Wang, Yue Zhang.
• Survey on Factuality in Large Language Models: Knowledge, Retrieval and Domain-Specificity [pdf][Github Page]
  Accepted by ACM Computing Surveys
  Cunxiang Wang*, Xiaoze Liu*, Yuanhao Yue*, Xiangru Tang, Tianhang Zhang, Cheng Jiayang, Yunzhi Yao, Wenyang Gao, Xuming Hu, Zehan Qi, Yidong Wang, Linyi Yang, Jindong Wang, Xing Xie, Zheng Zhang and Yue Zhang.
• StepMathAgent: A Step-Wise Agent for Evaluating Mathematical Processes through Tree-of-Error [pdf][Github Page]
  Shu-Xun Yang*, Cunxiang Wang* †, Yidong Wang, Xiaotao Gu, Minlie Huang, Jie Tang

  2024

• RedHat: Towards Reducing Hallucination in Essay Critiques with Large Language Models [pdf][Github Page]
  Andrew Zhuoer Feng, Cunxiang Wang*, Bosi Wen, Boqun Kou, Haohan Chi, Yu Luo, Lin Fan, Xiaotao Gu, Jie Tang, Hongning Wang, Minlie Huang
• Nash CoT: Multi-Path Inference with Preference Equilibrium [pdf][Github Page]
  EMNLP-2024
  Ziqi Zhang*, Cunxiang Wang*, Xiong Xiao, Yue Zhang, Donglin Wang

  2023

• Evaluating Open-QA Evaluation.
  Neurips-2023 Dataset and Benchmark Track. Average Score: 7.0. [pdf][codes and data]
  Cunxiang Wang*, Sirui Cheng*, Qipeng Guo, Zhikun Xu, Bowen Ding, Yunhao Yun, Yidong Wang, Xiangkun Hu, Zheng Zhang and Yue Zhang.
• Exploiting Abstract Meaning Representation for Open-Domain Question Answering. [pdf][codes and data].
  ACL-2023 findings, reviewer scores: 4, 4, 4 out of 5.
  Cunxiang Wang, Zhikun Xu, Qipeng Guo, Xiangkun Hu, Xuefeng Bai, Zheng Zhang and Yue Zhang.
• RFiD: Towards Rational Fusion-in-Decoder for Open-Domain Question Answering. [pdf][code and data].
  ACL-2023 findings, reviewer scores: 4, 4, 3 out of 5.
  Cunxiang Wang*, Haofei Yu* and Yue Zhang.
• Knowledgeable Salient Span Mask for Enhancing Language Models as Knowledge Base. [pdf][code and data].
  CCF International Conference on Natural Language Processing and Chinese Computing (NLPCC-2023).
  Cunxiang Wang, Fuli Luo, Yanyang Li, Runxin Xu, Fei Huang and Yue Zhang.
• TRAMS: Training-free Memory Selection for Long-range Language Modeling. [pdf][codes and data].
  EMNLP-2023 findings, reviewer scores: 4, 4, 3 out of 5.
  Haofei Yu*, Cunxiang Wang*, Yue Zhang and Wei Bi.

  2021

• Exploring Generalization Ability of Pretrained Language Models on Arithmetic and Logical Reasoning [pdf][data].
  CCF International Conference on Natural Language Processing and Chinese Computing (NLPCC-2021)
  Cunxiang Wang*, Boyuan Zheng*, Yuchen Niu and Yue Zhang.
• Can Generative Pre-trained Language Models Serve as Knowledge Bases for Closed-book QA? [pdf][codes and data].
  ACL-2021
  Cunxiang Wang*, Pai Liu*, Yue Zhang.

  2020

• SemEval-2020 task 4: Commonsense Validation and Explanation [pdf][codes and data].
  Proceedings of the 14th International Workshop on Semantic Evaluation (SemEval-2020).
  Cunxiang Wang*, Shuailong Liang*, Yili Jin, Yilong Wang, Xiaodan Zhu, Yue Zhang.
• Commonsense Knowledge Graph Reasoning by Selection or Generation? Why? arxiv [pdf].
  Cunxiang Wang, Jinhang Wu, Luxin Liu and Yue Zhang.

  2019

• Does it Make Sense? And Why? A Pilot Study for Sense Making and Explanation. [pdf][codes and data].
  ACL-2019
  Cunxiang Wang, Shuailong Liang, Yue Zhang, Xiaonan Li and Tian Gao.
• Domain Representation for Knowledge Graph Embedding. [pdf][codes and data].
  CCF International Conference on Natural Language Processing and Chinese Computing (NLPCC-2019).
  Cunxiang Wang, Feiliang Ren, Chenxu Zhao, ZhiChao Lin, Tian Xie and Yue Zhang.