Hey guys, this is 伍鸿秋 (Hongqiu Wu). I am now a PhD student in Deptment of CS, Shanghai Jiao Tong University, supervised by Professor Hai Zhao and Min Zhang. My research interest includes open virtual world, interactive drama, and error correction. I achieved the best on AdvGLUE and HellaSWAG. I’m serveing as reviewer/area chair for ARR, NeurIPS, etc.

I’m working as an intern at Anuttacon.

Project

Interactive Drama

Our second work explored the empowerment of the player that can have an impact on the characters in the drama, which is crucial for building a deep interactive experience. We call this Agency. [demo] [ACL 2025]

We developed LLM-based interactive drama in a script of Detective Conan, where audience can step into the play to interact with the characters. [repo] [zhihu] [ACL 2024]

This is our demonstration of LLM-based interactive Drama, script 《候车室的七个人》 (Seven in the Waiting Room).

Free Pokémon

A new game genre Open Role-Playing Game (ORPG). Players are allowed to craft their own pokemons by issuing natural language scripts. [repo] [video]

ReLM

We released ReLM (Rephrasing Language Model), a new state-of-the-art standard for Chinese Spelling Correction (CSC). [repo] [zhihu]; LEMON, a novel multi-domain CSC benchmark with ByteDance. [repo]

Education

• PhD, Department of Computer Science and Engineering, Shanghai Jiao Tong University, 2020 - present

• Bachelor of Engineering, Information Engineering, Shanghai Jiao Tong University, 2015 - 2020

Publication

* denotes equal contribution

2025

• [ACL 2025] Hongqiu Wu, Weiqi Wu, Tianyang Xu, Jiameng Zhang, Hai Zhao. Towards Enhanced Immersion and Agency for LLM-based Interactive Drama.
• [ACL 2025] Jiale Hong*, Hongqiu Wu*, Hai Zhao. Game Development as Human-LLM Interaction.
• [ACL 2025] Weiqi Wu, Hongqiu Wu, Hai Zhao. X-TURING: Towards an Enhanced and Efficient Turing Test for Long-Term Dialogue Agents.
• [COLING 2025] Linfeng Liu, Hongqiu Wu, Hai Zhao. Driving Chinese Spelling Correction from a Fine-Grained Perspective.

2024

• [EMNLP 2024 Demo] Hongqiu Wu*, Xingyuan Liu*, Yan Wang, Hai Zhao. Instruction-Driven Game Engine: A Poker Case Study.
• [Arxiv] Hongqiu Wu, Zekai Xu, Tianyang Xu, Jiale Hong, Weiqi Wu, Yan Wang, Hai Zhao, Min Zhang, Zhezhi He. Evolving Virtual World with Delta-Engine.
• [ACL 2024] Weiqi Wu*, Hongqiu Wu*, Lai Jiang, Xingyuan Liu, Jiale Hong, Hai Zhao, Min Zhang. From Role-Play to Drama-Interaction: An LLM Solution.
• [ACL 2024] Lai Jiang*, Hongqiu Wu*, Hai Zhao, Min Zhang. Chinese Spelling Corrector Is Just Language Learner.
• [COLING 2024] Yifei Yang*, Hongqiu Wu*, Hai Zhao. Attack Named Entity Recognition by Entity Boundary Interference.
• [COLING 2024] Khai Jiet Liong, Hongqiu Wu, Hai Zhao. Unveiling Vulnerability of Self-Attention.
• [AAAI 2024] Linfeng Liu*, Hongqiu Wu*, Hai Zhao. Chinese Spelling Correction as Rephrasing Language Model.

2023

• [EMNLP 2023] Hongqiu Wu, Linfeng Liu, Hai Zhao, Min Zhang. Empower Nested Boolean Logic via Self-Supervised Curriculum Learning.
• [ACL 2023] Hongqiu Wu, Shaohua Zhang, Yuchen Zhang, Hai Zhao. Rethinking Masked Language Modeling for Chinese Spelling Correction.
• [ICLR 2023] Hongqiu Wu, Yongxiang Liu, Hanwen Shi, Hai Zhao, Min Zhang. Toward Adversarial Training on Contextualized Language Representation.
• [AAAI 2023] Hongqiu Wu, Ruixue Ding, Hai Zhao, Pengjun Xie, Fei Huang, Min Zhang. Adversarial Self-Attention for Language Understanding.

2022

• [EMNLP 2022] Hongqiu Wu, Ruixue Ding, Hai Zhao, Boli Chen, Pengjun Xie, Fei Huang, Min Zhang. Forging Multiple Training Objectives for Pre-trained Language Models via Meta-Learning.
• [COLING 2022] Yiyang Li, Hongqiu Wu, Hai Zhao. Semantic-Preserving Adversarial Code Comprehension.

2021

• [Arxiv] Hongqiu Wu, Hai Zhao, Min Zhang. Not All Attention Is All You Need.
• [ACL 2021] Hongqiu Wu, Hai Zhao, Min Zhang. Code Summarization with Structure-induced Transformer.

Honor

• National Scholarship (国家奖学金)
• Huawei Scholarship (华为奖学金)
• Intel Scholarship (英特尔奖学金)
• Third place in Obei Cup
• Outstanding Student Scholarship
• The SJTU Cup Champions for Football

Talk

• LLM-based Interactive Drama, Netease, Hangzhou, 2024.11
• Open Virtual World and Delta-Engine, SJTU-CS Global Lunch Seminar, Shanghai, 2024.10
• LLM-based Interactive Drama, Minimax, Shanghai, 2024.9

Experience

• Research Intern, miHoYo LumiNLP
• Research Intern, ByteDance AI Lab
• Research Intern, Alibaba DAMO Academy
• Research Intern, Xiaomi AI Lab
• Algorithm Engineer Intern, Jing Dong Digits
• Algorithm Engineer Intern, Ping An Insurance