Papers by Wenbo Zhang
Chinese SafetyQA: A Safety Short-form Factuality Benchmark for Large Language Models (2025.acl-long)
Copied to clipboard
Yingshui Tan, Boren Zheng, Baihui Zheng, Kerui Cao, Huiyun Jing, Jincheng Wei, Jiaheng Liu, Yancheng He, Wenbo Su, Xiaoyong Zhu, Bo Zheng, Kaifu Zhang
| Challenge: | Large language models have created significant safety concerns . factuality ability is crucial in determining whether they can be deployed and applied safely and compliantly within specific regions. |
| Approach: | They propose a benchmark to evaluate the factuality of large language models in China . they evaluate the models' ability to provide accurate and reliable information . |
| Outcome: | The proposed benchmark evaluates the factuality abilities of existing LLMs and compares them to LLM abilities. |
E2-LLM: Efficient and Extreme Length Extension of Large Language Models (2024.findings-acl)
Copied to clipboard
Jiaheng Liu, ZhiqiBai ZhiqiBai, Yuanxing Zhang, Chenchen Zhang, YuangZh YuangZh, Ge Zhang, JiakaiWang JiakaiWang, Haoran Que, Yukang Chen, Wenbo Su, Tiezheng Ge, Jie Fu, Wenhu Chen, Bo Zheng
| Challenge: | Existing techniques for extending context capabilities in LLMs require additional training procedures and access to datasets with long context (e.g., sequences of 32K tokens). |
| Approach: | They propose a solution to extend context capabilities in Large Language Models by training a single process over a sequence of 4K tokens. |
| Outcome: | The proposed solution significantly reduces the cost of continual-pretraining or fine-tuning over short sequences and improves robustness to diverse relative positions. |
USB: A COMPREHENSIVE AND UNIFIED SAFETY EVALUATION BENCHMARK FOR MULTIMODAL LARGE LANGUAGE MODELS (2026.acl-long)
Copied to clipboard
Baolin Zheng, Guanlin Chen, Qingyang Teng, Hongqiong Zhong, Yingshui Tan, Zhendong Liu, Weixun Wang, Jiaheng Liu, Jian Yang, Huiyun Jing, Jincheng Wei, Wenbo Su, Xiaoyong Zhu, Bo Zheng, Kaifu Zhang
| Challenge: | Existing safety benchmarks fail to provide reliable assessments due to limited risk coverage, insufficient scale and the oversight of complex modality combinations. |
| Approach: | They propose a framework that covers 61 risk categories across four modality interactions to address this gap. |
| Outcome: | The proposed framework covers 61 risk categories across four distinct modality interactions. |
SELECting over Tokens: Curating Pre-training Data at Scale via Token Classification (2026.acl-long)
Copied to clipboard
Xin Tong, Weidong Zhang, Jiaang Li, Haibin Chen, Shilei Liu, Langming Liu, Kangtao Lv, Yujin Yuan, Wenbo Su, Bo Zheng
| Challenge: | Existing pipelines rely on expert-crafted heuristic rules, which lack content-aware, fine-grained noise detection. |
| Approach: | They propose a framework that reframes data refinement as a highly efficient token classification task. |
| Outcome: | The proposed framework outperforms existing pipelines on benchmarks and is 2.5x faster at inference. |
MES-RAG: Bringing Multi-modal, Entity-Storage, and Secure Enhancements to RAG (2025.findings-naacl)
Copied to clipboard
| Challenge: | Existing approaches to retrieve entity information are limited by document level retrieval and intermingled storage of information from different entities. |
| Approach: | They propose a framework that enhances entity-specific query handling . MES-RAG introduces proactive security measures that ensure system integrity . |
| Outcome: | Experimental results show that MES-RAG improves accuracy and recall . the framework can be integrated into existing RAG architectures . |
Rumor Detection on Social Media with Crowd Intelligence and ChatGPT-Assisted Networks (2023.emnlp-main)
Copied to clipboard
| Challenge: | Existing research on rumor detection challenges the expressive power of text encoding sequences, and insufficient mining of semantic structural information. |
| Approach: | They propose a Crowd Intelligence-based semantic feature learning module to capture textual content’s sequential and hierarchical features and a knowledge-based structural mining module that leverages ChatGPT for knowledge enhancement. |
| Outcome: | The proposed system achieves performance improvement in rumor detection tasks validating the effectiveness and rationality of using large language models as auxiliary tools. |
Recognizing Limits: Investigating Infeasibility in Large Language Models (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Large language models fail to handle queries that exceed their knowledge and capabilities, leading to incorrect or fabricated responses. |
| Approach: | They conceptualize four main categories of infeasible tasks for LLMs which cover a broad spectrum of hallucination-related challenges identified in prior literature. |
| Outcome: | The proposed models can handle requests exceeding their knowledge and capabilities and refuse them . |
MUCH: A Multimodal Corpus Construction for Conversational Humor Recognition Based on Chinese Sitcom (2024.lrec-main)
Copied to clipboard
| Challenge: | Existing multimodal corpora for conversational humor are coarse-grained and insufficient to support the conversational comprehension task. |
| Approach: | They constructed a multimodal humor corpus based on a Chinese sitcom and used both unimodal and multimodal methods to test the corpus. |
| Outcome: | The proposed method outperforms unimodal and multimodal methods in the evaluation of a Chinese sitcom for conversational humor recognition. |
Read As Human: Compressing Context via Parallelizable Close Reading and Skimming (2026.acl-long)
Copied to clipboard
Jiwei Tang, Shilei Liu, Zhicheng Zhang, Qingsong Lv, Runsong Zhao, Tingwei Lu, Langming Liu, Haibin Chen, Yujin Yuan, Hai-Tao Zheng, Wenbo Su, Bo Zheng
| Challenge: | Existing task-aware methods require loading the entire input sequence at once for compression, which suffer from computational inefficiency. |
| Approach: | They propose a framework that adopts an adaptive hybrid reading strategy to reduce computational inefficiency and redundant information in long-context scenarios. |
| Outcome: | Experiments show that RAM outperforms baselines on multiple question answering and summarization benchmarks while delivering up to a 12x speedup on long inputs. |
Have LLMs Reopened the Pandora’s Box of AI-Generated Fake News? (2025.naacl-long)
Copied to clipboard
Xinyu Wang, Wenbo Zhang, Sai Koneru, Hangzhi Guo, Bonam Mingole, S. Shyam Sundar, Sarah Rajtmajer, Amulya Yadav
| Challenge: | Large language models (LLMs) are increasingly being used by fake news creators to generate deceptive and persuasive content at scale. |
| Approach: | They propose to use large language models to generate fake news at scale and to assess the ability of human annotators and AI models to detect it. |
| Outcome: | The results show that LLMs are 68% more effective at detecting real news than humans, compared to humans and AI models for fake news detection. |
PLAWBENCH: A Rubric-Based Benchmark for Evaluating LLMs in Real-World Legal Practice (2026.acl-long)
Copied to clipboard
Yuzhen Shi, Huanghai Liu, Yiran HU, Song Gaojie, Xu Xinran, Yubo Ma, Tianyi Tang, Li Zhang, Qingjing Chen, Feng Di, Wenbo Lv, Weiheng Wu, Kexin Yang, Sen Yang, Wei Wang, Rongyao Shi, Qiu Yuanyang, Yuemeng Qi, Zhang Jingwen, Sui Xiaoyu, Yifan Chen, Zhang Yi, An Yang, Bowen Yu, Dayiheng Liu, Junyang Lin, Weixing Shen, Bing Zhao, Charles L. A. Clarke, HU Wei
| Challenge: | Existing benchmarks for large language models (LLMs) are coarse, single-dimensional metrics and do not explicitly assess fine-grained legal reasoning. |
| Approach: | They propose a Practical Law Benchmark to evaluate large language models in real-world legal practice scenarios. |
| Outcome: | The proposed model is based on 850 questions and 13 scenarios with expert-designed evaluation rubrics. |
Beyond the Singular: Revealing the Value of Multiple Generations in Benchmark Evaluation (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing evaluation methods for large language models overlook the inherent randomness of LLMs. |
| Approach: | They propose a hierarchical statistical model that incorporates both benchmark characteristics and LLM randomness to provide a more comprehensive representation of benchmarking process. |
| Outcome: | The proposed model improves the accuracy of estimating the benchmark score and reduces variance. |
A Quantum-Inspired Matching Network with Linguistic Theories for Metaphor Detection (2024.lrec-main)
Copied to clipboard
| Challenge: | Metaphor identification procedures and selectional preference violations are challenging for machines to recognize and comprehend metaphors. |
| Approach: | They propose a quantum-inspired matching network for metaphor detection based on QLM . metaphors are widely present in the language, thought and behavior of humans . |
| Outcome: | The proposed method can be used to detect metaphors even in the face of conventional metaphors. |
End-to-end Task-oriented Dialogue: A Survey of Tasks, Methods, and Future Directions (2023.emnlp-main)
Copied to clipboard
| Challenge: | End-to-end task-oriented dialogue (EToD) can generate responses in an end-to end fashion without modular training, which attracts escalating popularity. |
| Approach: | They present a systematic review of EToD and propose a unified perspective to summarize existing approaches and recent trends. |
| Outcome: | The proposed approaches can generate responses in an end-to-end fashion without modular training, which attracts escalating popularity. |
The Art of (Mis)alignment: How Fine-Tuning Methods Effectively Misalign and Realign LLMs in Post-Training (2026.findings-acl)
Copied to clipboard
Rui Zhang, Hongwei Li, Yun Shen, Xinyue Shen, Wenbo Jiang, Guowen Xu, Yang Liu, Michael Backes, Yang Zhang
| Challenge: | Misaligned large language models can magnify harm by exploiting them to undermine safety . et al., 2022b; Bai e.t., 2023): misalignment, realignment and model-specific resistance are important . |
| Approach: | They evaluate four methods to identify a mechanism asymmetry between attack and defense . they find that ORPO is most effective for misalignment, but DPO excels in realignment . |
| Outcome: | The proposed methods show a mechanism asymmetry between attack and defense . the proposed methods excel in realignment, but at the expense of model utility . |
RAVR: Reference-Answer-guided Variational Reasoning for Large Language Models (2026.findings-acl)
Copied to clipboard
| Challenge: | Experiments show that reinforcement learning (RL) can refine the reasoning abilities of large language models (LLMs) but requires a key prerequisite: the model must already be able to generate high-utility reasoning paths with non-negligible probability. |
| Approach: | They propose a framework that uses answer-conditioned reasoning as a variational surrogate for question-only reasoning. |
| Outcome: | Experiments on 11 benchmarks and 3 models show that RAVR reduces hesitation, strengthens conclusion consolidation, and promotes problem-specific strategies in reasoning. |
CoMeT: Collaborative Memory Transformer for Efficient Long Context Modeling (2026.acl-long)
Copied to clipboard
Runsong Zhao, Shilei Liu, Jiwei Tang, Langming Liu, Haibin Chen, Weidong Zhang, Yujin Yuan, Tong Xiao, JingBo Zhu, Wenbo Su, Bo Zheng
| Challenge: | a novel architecture that enables LLMs to handle arbitrarily long sequences with constant memory usage and linear time complexity is a major barrier to long-context processing. |
| Approach: | They propose a novel architecture that enables LLMs to handle arbitrarily long sequences with constant memory usage and linear time complexity. |
| Outcome: | The proposed architecture can handle arbitrarily long sequences with constant memory usage and linear time complexity. |
A Functionality-Grounded Benchmark for Evaluating Web Agents in E-commerce Domains (2026.acl-long)
Copied to clipboard
| Challenge: | Existing benchmarks focus on product search tasks, but ignore potential risks. |
| Approach: | They propose a data generation pipeline that leverages webpage content and interactive elements to create diverse, functionality-grounded user queries. |
| Outcome: | The proposed framework assesses the performance and safety of web agents under dynamic, real-world e-commerce environments. |
PretrainRL: Alleviating Factuality Hallucination of Large Language Models at the Beginning (2026.findings-acl)
Copied to clipboard
Langming Liu, Kangtao Lv, Haibin Chen, Weidong Zhang, Yejing Wang, Shilei Liu, Xin Tong, Yujin Yuan, Yongwei Wang, Wenbo Su, Bo Zheng
| Challenge: | Large language models suffer from factual hallucinations where they generate verifiable falsehoods. |
| Approach: | They propose a framework that integrates reinforcement learning into the pretraining phase to consolidate factual knowledge. |
| Outcome: | The proposed framework significantly alleviates factual hallucinations and outperforms state-of-the-art methods. |
ConceptMath: A Bilingual Concept-wise Benchmark for Measuring Mathematical Reasoning of Large Language Models (2024.findings-acl)
Copied to clipboard
Yanan Wu, Jie Liu, Xingyuan Bu, Jiaheng Liu, Zhanhui Zhou, Yuanxing Zhang, Chenchen Zhang, ZhiqiBai ZhiqiBai, Haibin Chen, Tiezheng Ge, Wanli Ouyang, Wenbo Su, Bo Zheng
| Challenge: | ConceptMath evaluates concept-wise mathematical reasoning of Large Language Models (LLMs) Existing benchmarks that evaluate general mathematical reasoning with an average accuracy fail to probe the fine-grained failure modes of mathematical reasoning on specific datasets. |
| Approach: | They introduce a bilingual, fine-grained benchmark that evaluates concept-wise mathematical reasoning of Large Language Models. |
| Outcome: | The proposed benchmarks evaluate concept-wise mathematical reasoning of Large Language Models with concept-based accuracies. |
Incorporating Object-Level Visual Context for Multimodal Fine-Grained Entity Typing (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Experimental results show that fine-grained entity typing is superior to text-based methods. |
| Approach: | They propose a task called fine-grained entity typing to classify entities . they propose combining textual and visual contexts to capture fine-granular semantic information . |
| Outcome: | The proposed approach achieves superior classification performance compared to previous text-based approaches. |
M2RC-EVAL: Massively Multilingual Repository-level Code Completion Evaluation (2025.acl-long)
Copied to clipboard
Jiaheng Liu, Ken Deng, Congnan Liu, Jian Yang, Shukai Liu, He Zhu, Peng Zhao, Linzheng Chai, Yanan Wu, JinKe JinKe, Ge Zhang, Zekun Moore Wang, Guoan Zhang, Yingshui Tan, Bangyu Xiang, Zhaoxiang Zhang, Wenbo Su, Bo Zheng
| Challenge: | Existing repository-level code completion benchmarks focus on a limited number of languages . existing benchmarks report overall average scores of different languages ignoring fine-grained abilities . |
| Approach: | They propose to use repository-level code completion benchmarks to evaluate general code intelligence abilities across languages for existing code Large Language Models. |
| Outcome: | The proposed benchmarks improve the code completion abilities of existing LLMs by using two types of annotations on the parsed syntax tree. |
NavA3: Understanding Any Instruction, Navigating Anywhere, Finding Anything (2026.acl-long)
Copied to clipboard
Lingfeng Zhang, Xiaoshuai Hao, Yingbo Tang, Haoxiang Fu, Xinyu Zheng, Pengwei Wang, Zhongyuan Wang, Wenbo Ding, Shanghang Zhang
| Challenge: | Existing embodied navigation methods struggle with such tasks due to their limitations in comprehending high-level human instructions and localizing objects with an open vocabulary. |
| Approach: | They propose a hierarchical framework for long-horizon navigation that integrates human instructions with 3D scene views. |
| Outcome: | The proposed model achieves SOTA results and can complete long-horizon navigation tasks across different robot embodiments in real-world environments. |
Can Large Language Models Detect Errors in Long Chain-of-Thought Reasoning? (2025.acl-long)
Copied to clipboard
Yancheng He, Shilong Li, Jiaheng Liu, Weixun Wang, Xingyuan Bu, Ge Zhang, Z.y. Peng, Zhaoxiang Zhang, Zhicheng Zheng, Wenbo Su, Bo Zheng
| Challenge: | Recent advances in o1-like models have generated long Chain-of-Thought reasoning steps to improve the reasoning abilities of existing Large Language Models (LLMs). |
| Approach: | They propose a DeltaBench to analyze the quality and effectiveness of o1-like models and measure their ability to detect errors in long COT reasoning. |
| Outcome: | The proposed model can detect errors in long COT reasoning. |
Evaluating Robustness of Large Audio Language Models to Audio Injection: An Empirical Study (2025.emnlp-main)
Copied to clipboard
| Challenge: | Large Audio-Language Models (LALMs) are increasingly being deployed in real-world applications, yet their robustness against malicious audio injection remains underexplored. |
| Approach: | They quantitatively assess their vulnerabilities and resilience using metrics: the Defense Success Rate, Context Robustness Score, and Judgment Robustic Index. |
| Outcome: | The proposed models demonstrate significant performance disparities across four attack scenarios. |