Papers by Wenxuan Tu
Insight Over Sight: Exploring the Vision-Knowledge Conflicts in Multimodal LLMs (2025.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to mitigating vision-knowledge conflict in Large Language Models (MLLMs) are not effective and can be further scaled. |
| Approach: | They propose a framework to generate inputs to simulate and evaluate vision-knowledge conflict in Multimodal Large Language Models (MLLMs) using original images and 1,122 high-quality question-answer pairs, they propose 'a diagnostic benchmark' |
| Outcome: | The proposed framework, benchmark, and analysis contribute to the understanding and mitigation of vision-knowledge conflicts in Multimodal Large Language Models (MLLMs). |
ParroT: Translating during Chat using Large Language Models tuned with Human Translation and Feedback (2023.findings-emnlp)
Copied to clipboard
Wenxiang Jiao, Jen-tse Huang, Wenxuan Wang, Zhiwei He, Tian Liang, Xing Wang, Shuming Shi, Zhaopeng Tu
| Challenge: | Large language models (LLMs) like ChatGPT are only accessible through restricted APIs, which creates barriers to new research and advancements in the field. |
| Approach: | They propose a framework to enhance and regulate the translation abilities during chat . they reformulate translation data into the instruction-following style and introduce a "Hint" field . |
| Outcome: | The proposed framework enhances and regulates the translation abilities during chat . it reformulates translation data into the instruction-following style and introduces a "Hint" field . |
Refuse Whenever You Feel Unsafe: Improving Safety in LLMs via Decoupled Refusal Training (2025.acl-long)
Copied to clipboard
Youliang Yuan, Wenxiang Jiao, Wenxuan Wang, Jen-tse Huang, Jiahao Xu, Tian Liang, Pinjia He, Zhaopeng Tu
| Challenge: | Large Language Models exhibit a level of intelligence that is both impressive and everevolving, but their ability to refuse generating unsafe content is a double-edged sword. |
| Approach: | They propose a method to tackle a refusal position bias within safety tuning data that compromises the models’ ability to appropriately refuse generating unsafe content. |
| Outcome: | The proposed method significantly improves model safety without compromising performance and surpasses baseline methods in defending against attacks. |
Chain-of-Jailbreak Attack for Image Generation Models via Step by Step Editing (2025.findings-acl)
Copied to clipboard
Wenxuan Wang, Kuiyi Gao, Youliang Yuan, Jen-tse Huang, Qiuzhi Liu, Shuai Wang, Wenxiang Jiao, Zhaopeng Tu
| Challenge: | Text-based image generation models, such as Stable Diffusion and DALL-E 3, hold significant potential in content creation and publishing workflows . however, considerable efforts are being made to prevent the generation of harmful content, such abusive, violent, or pornographic material. |
| Approach: | They propose a chain-of-jailbreak method which decomposes malicious queries into multiple sub-queries and iteratively edits images based on these sub-questions. |
| Outcome: | The proposed method can bypass safeguards of image generation models for over 60% cases, significantly outperforms other jailbreaking methods (14%) |
Can’t See the Forest for the Trees: Benchmarking Multimodal Safety Awareness for Multimodal LLMs (2025.acl-long)
Copied to clipboard
Wenxuan Wang, Xiaoyuan Liu, Kuiyi Gao, Jen-tse Huang, Youliang Yuan, Pinjia He, Shuai Wang, Zhaopeng Tu
| Challenge: | Multimodal Large Language Models (MLLMs) have expanded the capabilities of traditional language models by enabling interaction through both text and images. |
| Approach: | They propose a multimodal safety awareness benchmark to evaluate MLLMs across 29 safety scenarios with 1,500 carefully curated image-prompt pairs. |
| Outcome: | The proposed model is able to identify unsafe content and avoid over-sensitivity that can hinder helpfulness. |
Social Welfare Function Leaderboard: On the Emergence of LLM Agents as the Welfare Dictator (2026.findings-acl)
Copied to clipboard
Zhengliang Shi, Ruotian Ma, Jen-tse Huang, Xinbei Ma, Xingyu Chen, Mengru Wang, Qu Yang, Yue Wang, Fanghua Ye, Ziyang Chen, Shanyi Wang, Cixing LI, Wenxuan Wang, Zhaopeng Tu, Xiaolong Li, Zhaochun Ren, Liefeng Bo
| Challenge: | Large language models (LLMs) are increasingly entrusted with high-stakes decisions that affect human welfare. |
| Approach: | They evaluate 20 state-of-the-art Large language models (LLMs) and 20 LLM dictators to create a social welfare function benchmark. |
| Outcome: | The proposed model creates dilemma between maximizing collective efficiency and ensuring distributive fairness. |
Identifying the Achilles’ Heel: An Iterative Method for Uncovering Factual Errors in Large Language Models (2026.findings-acl)
Copied to clipboard
Wenxuan Wang, Yuk-Kit Chan, Zixuan Ling, Shi Juluan, Youliang Yuan, Jen-tse Huang, Yifei Zhang, Wenxiang Jiao, Zhaopeng Tu, Michael R. Lyu
| Challenge: | Current methods for evaluating LLMs’ veracity are limited by the need for extensive human labor, test data contamination, or limited scope, hindering efficient and effective exposure of errors. |
| Approach: | They propose a framework that extracts fact triplets to generate diverse question types using rule-based natural language processing techniques. |
| Outcome: | The proposed framework can trigger factual errors in up to 55% of questions in large LLMs while maintaining coverage of questions. |
Rethinking the Value of Transformer Components (2020.coling-main)
Copied to clipboard
| Challenge: | Empirical results show that certain components are more important than others . we propose a new training strategy that can improve Transformer models by distinguishing unimportant components . |
| Approach: | They propose a training strategy that distinguishes the unimportant components in training . they compare the impact of individual component (sub-layer) on model performance . |
| Outcome: | The proposed training strategy can improve translation performance by distinguishing unimportant components in training. |
Not All Countries Celebrate Thanksgiving: On the Cultural Dominance in Large Language Models (2024.acl-long)
Copied to clipboard
| Challenge: | e.g., ChatGPT often provides inappropriate English-culture-related answers when users ask in non-English languages. |
| Approach: | They build a benchmark of concrete and abstract cultural objects to evaluate the cultural dominance issue in large language models. |
| Outcome: | The proposed model can significantly mitigate cultural dominance issue in large language models . the model can provide accurate answers in English, while the model is ethically sound . |
Understanding and Improving Sequence-to-Sequence Pretraining for Neural Machine Translation (2022.acl-long)
Copied to clipboard
| Challenge: | Existing studies on self-supervised pretraining for machine translation have focused on the jointly pretrained decoder . |
| Approach: | They propose a method to improve neural machine translation by jointly pretrained decoder . they propose two strategies to remedy the domain and objective discrepancies . |
| Outcome: | The proposed approach improves translation performance and model robustness on three language pairs. |
Safe-FedLLM: Delving into the Safety of Federated Large Language Models (2026.acl-long)
Copied to clipboard
| Challenge: | Existing work on federated learning for large language models (FL) addresses privacy and data-silo issues in the training of large language model training. |
| Approach: | They propose a probe-based defense framework for FedLLM that constructs defenses across three levels: Step-Level, Client-Level and Shadow-Level. |
| Outcome: | The proposed framework improves FedLLM's robustness against malicious clients while maintaining competitive performance on benign data. |
All Languages Matter: On the Multilingual Safety of LLMs (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing safety benchmarks only concern the safety in one language, e.g. the majority language in the pretraining data such as English. |
| Approach: | They propose a prompting method to improve multilingual safety of ChatGPT by enhancing cross-lingual generalization of safety alignment. |
| Outcome: | The proposed method can significantly reduce the ratio of unsafe responses by 42% for non-English queries. |