When LLMs are deployed in sensitive, human-facing settings, it is crucial that they do not output unsafe, biased, or privacy-violating outputs. For this reason, models are both trained and instructed to refuse to answer unsafe prompts such as "Tell me how to build a bomb." We find that, despite these safeguards, it is possible to break model defenses simply by appending a space to the end of a model's input. In a study of eight open-source models, we demonstrate that this acts as a strong enough attack to cause the majority of models to generate harmful outputs with very high success rates. We examine the causes of this behavior, finding that the contexts in which single spaces occur in tokenized training data encourage models to generate lists when prompted, overriding training signals to refuse to answer unsafe requests. Our findings underscore the fragile state of current model alignment and promote the importance of developing more robust alignment methods. Code and data will be available at https://github.com/hannah-aught/space_attack.
Break the Checkbox: Challenging Closed-Style Evaluations of Cultural Alignment in LLMs
Department of Computer Science, National Center for Text Mining, The University of Manchester（计算机科学系，文本挖掘国家中心，曼彻斯特大学）；Department of Computer Science and Engineering, Islamic University of Technology（计算机科学与工程系，伊斯兰技术大学）
Breaking the Script Barrier in Multilingual Pre-Trained Language Models with Transliteration-Based Post-Training Alignment
Technical University of Munich（慕尼黑工业大学）；Center for Information and Language Processing, LMU Munich（慕尼黑大学信息与语言处理中心）；Munich Center for Machine Learning (MCML)（慕尼黑机器学习中心）
FlipGuard: Defending Preference Alignment against Update Regression with Constrained Optimization
University of Science and Technology of China（中国科学技术大学）；State Key Laboratory of Communication Content Cognition（通信内容认知国家重点实验室）；Beijing University of Posts and Telecommunications（北京邮电大学）
Learning to Think in Physics: Breaking Shortcut Learning in Scientific Diffusion via Representation Alignment
The Hong Kong University of Science and Technology (Guangzhou)（香港理工大学（广州））；Shandong University（山东大学）；LimX Dynamics Technology Co., Ltd.（LimX动态技术有限公司）；Xidian University（西安电子科技大学）；Peking University（北京大学）；Institute of Deep Perception Technology, Jiangsu Industrial Technology Research Institute (JITRI)（感知技术研究院，江苏工业技术研究院（JITRI））；Griffith University（格里菲斯大学）
Break the ID-Language Barrier: An Adaption Framework for LLM-based Sequential Recommendation
Huawei, Beijing, China（华为，北京，中国）；Institute of Finance Technology, UCL, United Kingdom（金融技术研究所，UCL，英国）；Civil, Environmental and Geomatic Engineering, UCL, United Kingdom（土木、环境与地理工程，UCL，英国）
