Papers by Xingyu Sui
Beware of Your Po! Measuring and Mitigating AI Safety Risks in Role-Play Fine-Tuning of LLMs (2025.acl-long)
Copied to clipboard
Weixiang Zhao, Yulin Hu, Yang Deng, Jiahe Guo, Xingyu Sui, Xinyang Han, An Zhang, Yanyan Zhao, Bing Qin, Tat-Seng Chua, Ting Liu
| Challenge: | Existing role-play fine-tuning techniques improve role adaptability but may degrade safety performance, especially for villainous characters. |
| Approach: | They propose safety-aware Role-Play Fine-Tuning (SaRFT) to balance role-playing capabilities and safety. |
| Outcome: | The proposed method outperforms state-of-the-art baselines under both LoRA and full-parameter fine-tuning settings. |
When Personalization Legitimizes Risks: Uncovering Safety Vulnerabilities in Personalized Dialogue Agents (2026.acl-long)
Copied to clipboard
Jiahe Guo, Xiangran Guo, Yulin Hu, Zimo Long, Xingyu Sui, Xuda Zhi, Yongbo Huang, Hao He, Weixiang Zhao, Yanyan Zhao, Bing Qin
| Challenge: | Existing research on personalized LLM agents focuses on the effectiveness of personalized responses. |
| Approach: | They propose a benchmark to quantify intent legitimation in personalized interactions . they propose 'detection-reflection' method that detects intent legititimation from internal representation space . |
| Outcome: | The proposed method reduces safety degradation by using internal representation space. |
Chain of Strategy Optimization Makes Large Language Models Better Emotional Supporter (2025.findings-emnlp)
Copied to clipboard
Weixiang Zhao, Xingyu Sui, Xinyang Han, Yang Deng, Yulin Hu, Jiahe Guo, Libo Qin, Qianyun Du, Shijin Wang, Yanyan Zhao, Bing Qin, Ting Liu
| Challenge: | Existing supervised fine-tuning (SFT) fails to address these issues, as it trains models on single gold-standard responses without modeling nuanced strategy trade-offs. |
| Approach: | They propose a two-stage framework that optimizes strategy selection preferences at each dialogue turn. |
| Outcome: | The proposed framework improves strategy selection preferences at each dialogue turn. |
AdaSteer: Your Aligned LLM is Inherently an Adaptive Jailbreak Defender (2025.emnlp-main)
Copied to clipboard
Weixiang Zhao, Jiahe Guo, Yulin Hu, Yang Deng, An Zhang, Xingyu Sui, Xinyang Han, Yanyan Zhao, Bing Qin, Tat-Seng Chua, Ting Liu
| Challenge: | Activation steering offers training-free defense but relies on fixed steering coefficients, resulting in suboptimal protection and increased false rejections of benign inputs. |
| Approach: | They propose an adaptive activation steering method that dynamically adjusts model behavior based on input characteristics. |
| Outcome: | The proposed method outperforms baseline methods across multiple jailbreak attacks with minimal impact on utility. |
TEA-Bench: A Systematic Benchmarking of Tool-enhanced Emotional Support Dialogue Agent (2026.acl-long)
Copied to clipboard
| Challenge: | Existing ESC systems focus on affective support in text-only settings, ignoring how external tools can enable factual grounding and reduce hallucination in multi-turn emotional support. |
| Approach: | They propose a benchmark for evaluating tool-augmented agents in ESC with realistic emotional scenarios and an MCP-style tool environment. |
| Outcome: | The proposed benchmarks show that tool augmentation improves emotional support quality and reduces hallucination, but weaker models benefit only marginally. |