Zixiang Xu's Website

Findings of EMNLP 2026 SocialSim @ COLM 2025 · Spotlight

SocialMaze: A Benchmark for Evaluating and Enhancing Social Reasoning in Large Language Models in Complex Social Environments

Zixiang Xu, Yanbo Wang, Yue Huang, Haomin Zhuang, Yujun Zhou, Jiayi Ye, et al.

Project Page Paper arXiv Code Data

ACL 2025

Cross-Lingual Pitfalls: Automatic Probing Cross-Lingual Weakness of Multilingual Large Language Models

Zixiang Xu*, Yanbo Wang*, Yue Huang*, Xiuying Chen, Jieyu Zhao, Meng Jiang, Xiangliang Zhang (* Equal contribution)

Project Page Paper Code

NeurIPS 2025

Adaptive Distraction: Probing LLM Contextual Robustness with Automated Tree Search

Yanbo Wang*, Zixiang Xu*, Yue Huang*, Chujie Gao, Siyuan Wu, Jiayi Ye, et al. (* Equal contribution)

Paper Code

NeurIPS 2025

DyFlow: Dynamic Workflow Framework for Agentic Reasoning

Yanbo Wang, Zixiang Xu, Yue Huang, Xiangqi Wang, Zirui Song, Lang Gao, Chenxi Wang, et al.

Paper Code

Preprint Under review

Inside the Unfair Judge: A Mechanistic Interpretability Account of LLM-as-Judge Bias

Zixiang Xu, Sixian Li, Huaxing Liu, Xiang Wang, Shuai Li, Zirui Song, Xiuying Chen

Project Page Paper

Preprint 2026

GTA: Graph Theory Agent and Benchmark for Algorithmic Graph Reasoning with LLMs

Zixiang Xu, Yanbo Wang, Chenxi Wang, Lang Gao, Zirui Song, Yue Huang, Zhaorun Chen, Xiangliang Zhang, Xiuying Chen

Paper Code

ICML 2026

Do LLMs “Feel”? Emotion Circuits Discovery and Control

Chenxi Wang, Yixuan Zhang, et al., Zixiang Xu, et al.

Paper Code

ACL 2026

Audio Jailbreak: An Open Comprehensive Benchmark for Jailbreaking Large Audio-Language Models

Zirui Song, Qian Jiang, et al., Zixiang Xu, et al.

Paper Code

AAAI 2026

ManipLVM-R1: Reinforcement Learning for Reasoning in Embodied Manipulation with Large Vision-Language Models

Zirui Song, Guangxian Ouyang, et al., Zixiang Xu, et al.

Paper

TRB 2025 Oral

Temporal-IRL: Modeling Port Congestion and Berth Scheduling with Inverse Reinforcement Learning

Yikuan Hu*, Zixiang Xu*, Wei Zhang*, et al. (* Equal contribution)

Paper

Findings of EMNLP 2025

Under the Shadow of Babel: How Language Shapes Reasoning in LLMs

Chenxi Wang, Yixuan Zhang, Lang Gao, Zixiang Xu, et al.

Paper

ACM MM 2024

GenUDC: High-Quality 3D Mesh Generation with Unsigned Dual Contouring Representation

Ruowei Wang, Jiaqi Li, et al., Zixiang Xu, et al.

Paper

arXiv 2025

Beyond Survival: Evaluating LLMs in Social Deduction Games with Human-Aligned Strategies

Zirui Song, Yuan Huang, Junchang Liu, et al., Zixiang Xu, et al.

Paper

arXiv 2025

Evaluate Bias without Manual Test Sets: A Concept Representation Perspective for LLMs

Lang Gao, Kaiyang Wan, Wei Liu, et al., Zixiang Xu, et al.

Paper

PRCV 2024 Oral

MTFusion: Reconstructing Any 3D Object from Single Image Using Multi-word Textual Inversion

Yu Liu, Ruowei Wang, Jiaqi Li, Zixiang Xu, Qijun Zhao

Paper

CCBR 2023 Oral

Text-Based Face Retrieval: Methods and Challenges

Yuchuan Deng, Qijun Zhao, Zhanpeng Hu, Zixiang Xu

Paper

Zixiang Xu（徐子翔）

Publications

SocialMaze: A Benchmark for Evaluating and Enhancing Social Reasoning in Large Language Models in Complex Social Environments

Cross-Lingual Pitfalls: Automatic Probing Cross-Lingual Weakness of Multilingual Large Language Models

Adaptive Distraction: Probing LLM Contextual Robustness with Automated Tree Search

DyFlow: Dynamic Workflow Framework for Agentic Reasoning

Inside the Unfair Judge: A Mechanistic Interpretability Account of LLM-as-Judge Bias

GTA: Graph Theory Agent and Benchmark for Algorithmic Graph Reasoning with LLMs

Do LLMs “Feel”? Emotion Circuits Discovery and Control

Audio Jailbreak: An Open Comprehensive Benchmark for Jailbreaking Large Audio-Language Models

ManipLVM-R1: Reinforcement Learning for Reasoning in Embodied Manipulation with Large Vision-Language Models

Temporal-IRL: Modeling Port Congestion and Berth Scheduling with Inverse Reinforcement Learning

Under the Shadow of Babel: How Language Shapes Reasoning in LLMs

GenUDC: High-Quality 3D Mesh Generation with Unsigned Dual Contouring Representation

Beyond Survival: Evaluating LLMs in Social Deduction Games with Human-Aligned Strategies

Evaluate Bias without Manual Test Sets: A Concept Representation Perspective for LLMs

MTFusion: Reconstructing Any 3D Object from Single Image Using Multi-word Textual Inversion

Text-Based Face Retrieval: Methods and Challenges