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. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Yancheng He, Shilong Li, Jiaheng Liu, Yingshui Tan, Weixun Wang, Hui Huang, Xingyuan Bu, Hangyu Guo, Chengwei Hu, Boren Zheng, Zhuoran Lin, Dekai Sun, Zhicheng Zheng, Wenbo Su, Bo Zheng. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

The integration of additional modalities increases the susceptibility of large vision-language models (LVLMs) to safety risks, such as jailbreak attacks, compared to their language-only counterparts. While existing research primarily focuses on post-hoc alignment techniques, the underlying safety mechanisms within LVLMs remain largely unexplored. In this work , we investigate whether LVLMs inherently encode safety-relevant signals within their internal activations during inference. Our findings reveal that LVLMs exhibit distinct activation patterns when processing unsafe prompts, which can be leveraged to detect and mitigate adversarial inputs without requiring extensive fine-tuning. Building on this insight, we introduce HiddenDetect, a novel tuning-free framework that harnesses internal model activations to enhance safety. Experimental results show that HiddenDetect surpasses state-of-the-art methods in detecting jailbreak attacks against LVLMs. By utilizing intrinsic safety-aware patterns, our method provides an efficient and scalable solution for strengthening LVLM robustness against multimodal threats. Our code and data will be released publicly.

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. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Yingshui Tan, Boren Zheng, Baihui Zheng, Kerui Cao, Huiyun Jing, Jincheng Wei, Jiaheng Liu, Yancheng He, Wenbo Su, Xiaoyong Zhu, Bo Zheng, Kaifu Zhang. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Large Language Models (LLMs) have showcased remarkable potential across various domains, especially in text generation. However, their vulnerability to jailbreak attacks presents considerable challenges to secure deployment, as attackers can use carefully crafted prompts to bypass safety measures and generate harmful content. Current jailbreak methods generally suffer from two significant limitations: a restricted strategy space for generating adversarial prompts and insufficient optimization of prompts based on feedback from LLMs. To overcome these challenges, we present Multistep Adaptive Attack Agent (MATA), an approach that employs a game-theoretic interaction between attack model and target model to adaptively execute jailbreak attacks on LLMs. This method enables iterative attempts based on reflection, gradually identifying the optimal jailbreak attack strategy within a complex strategy space. We compared MATA with mainstream methods across multiple open-source and closed-source LLMs, including Llama3.1, GLM4, and GPT4o. The results demonstrate that our approach exceeds existing methods in terms of attack success rate, average number of queries, and prompt diversity, effectively identifying vulnerabilities in LLMs.

Recent advancements in large language models (LLMs) underscore the need for more comprehensive evaluation methods to accurately assess their reasoning capabilities. Existing benchmarks are often domain-specific and thus cannot fully capture an LLM’s general reasoning potential. To address this limitation, we introduce the **Knowledge Orthogonal Reasoning Gymnasium (KORGym)**, a dynamic evaluation platform inspired by KOR-Bench and Gymnasium. KORGym offers over fifty games in either textual or visual formats and supports interactive, multi-turn assessments with reinforcement learning scenarios. Using KORGym, we conduct extensive experiments on 19 LLMs and 8 VLMs, revealing consistent reasoning patterns within model families and demonstrating the superior performance of closed-source models. Further analysis examines the effects of modality, reasoning strategies, reinforcement learning techniques, and response length on model performance. We expect KORGym to become a valuable resource for advancing LLM reasoning research and developing evaluation methodologies suited to complex, interactive environments.

As Visual Language Models (VLMs) continue to evolve, they have demonstrated increasingly sophisticated logical reasoning capabilities and multimodal thought generation, opening doors to widespread applications. However, this advancement raises serious concerns about content security, particularly when these models process complex multimodal inputs requiring intricate reasoning. When faced with these safety challenges, the critical competition between logical reasoning and safety objectives of VLMs is often overlooked in previous works. In this paper, we introduce Visualization-of-Thought Attack (\textbf{VoTA}), a novel and automated attack framework that strategically constructs chains of images with risky visual thoughts to challenge victim models. Our attack provokes the inherent conflict between the model's logical processing and safety protocols, ultimately leading to the generation of unsafe content. Through comprehensive experiments, VoTA achieves remarkable effectiveness, improving the average attack success rate (ASR) by 26.71\% (from 63.70\% to 90.41\%) on 9 open-source and 6 commercial VLMs, compared to the state-of-the-art methods. These results expose a critical vulnerability: current VLMs struggle to maintain safety guarantees when processing insecure multimodal visualization-of-thought inputs, highlighting the urgency and necessity of enhancing safety alignment. Our code and dataset are available at https://github.com/Hongqiong12/VoTA. Content Warning: This paper contains harmful contents that may be offensive.

We present a novel unsupervised deep learning approach that utilizes an encoder-decoder architecture for detecting anomalies in sequential sensor data collected during industrial manufacturing. Our approach is designed to not only detect whether there exists an anomaly at a given time step, but also to predict what will happen next in the (sequential) process. We demonstrate our approach on a dataset collected from a real-world Additive Manufacturing (AM) testbed. The dataset contains infrared (IR) images collected under both normal conditions and synthetic anomalies. We show that our encoder-decoder model is able to identify the injected anomalies in a modern AM manufacturing process in an unsupervised fashion. In addition, our approach also gives hints about the temperature non-uniformity of the testbed during manufacturing, which was not previously known prior to the experiment.