Large Language Models (LLMs) face a fundamental safety-helpfulness trade-off due to static, one-size-fits-all safety policies that lack runtime controllabilityxf, making it difficult to tailor responses to diverse application needs. %As a result, models may over-refuse benign requests or under-constrain harmful ones. We present \textbf{PACT} (Prompt-configured Action via Chain-of-Thought), a framework for dynamic safety control through explicit, risk-aware reasoning. PACT operates under a hierarchical policy architecture: a non-overridable global safety policy establishes immutable boundaries for critical risks (e.g., child safety, violent extremism), while user-defined policies can introduce domain-specific (non-global) risk categories and specify label-to-action behaviors to improve utility in real-world deployment settings. The framework decomposes safety decisions into structured Classify$\rightarrow$Act paths that route queries to the appropriate action (comply, guide, or reject) and render the decision-making process transparent.
Extensive experiments demonstrate that PACT achieves near state-of-the-art safety performance under global policy evaluation while attaining the best controllability under user-specific policy evaluation, effectively mitigating the safety-helpfulness trade-off. We will release the PACT model suite, training data, and evaluation protocols to facilitate reproducible research in controllable safety alignment.
Cognitive Loop of Thought: Reversible Hierarchical Markov Chain for Efficient Mathematical Reasoning
School of Computer Science and Technology, East China Normal University(1 计算机科学与技术学院,华东师范大学);School of Electronic and Electrical Engineering, Shanghai University of Engineering Science(2 电子电气工程学院,上海工程技术大学)
H-CoT: Hijacking the Chain-of-Thought Safety Reasoning Mechanism to Jailbreak Large Reasoning Models, Including OpenAI o1/o3, DeepSeek-R1, and Gemini 2.0 Flash Thinking
Duke University(杜克大学);Accenture(埃森哲);Accenture Security(埃森哲安全)
AdvChain: Adversarial Chain-of-Thought Tuning for Robust Safety Alignment of Large Reasoning Models
The Chinese University of Hong Kong, Shenzhen(香港中文大学(深圳));State University of New York at Buffalo(纽约州立大学布法罗分校);Huawei International, Singapore(新加坡华为国际)
HGOT: Hierarchical Graph of Thoughts for Retrieval-Augmented In-Context Learning in Factuality Evaluation
Smith School of Business, Queen’s University(女王大学史密斯商学院);Ingenuity Labs Research Institute, Queen’s University(女王大学 Ingenuity 实验室研究所);Department of Electrical and Computer Engineering, Queen’s University(女王大学电气与计算机工程系)
ConsisGuard: Aligning Safety Deliberation with Policy Enforcement in LLM Guardrails
Alibaba Group(阿里巴巴集团);Zhejiang University(浙江大学);Huzhou Normal University(湖州师范学院);Zhejiang Key Laboratory of Intelligent Education Technology and Application(浙江省智能教育技术与应用重点实验室)
SALAD-Bench: A Hierarchical and Comprehensive Safety Benchmark for Large Language Models
Shanghai Artificial Intelligence Laboratory(上海人工智能实验室);Harbin Institute of Technology(哈尔滨工业大学);Chinese University of Hong Kong(香港中文大学);Beijing Institute of Technology(北京理工大学);The Hong Kong Polytechnic University(香港理工大学)
SafetyAnalyst: Interpretable, Transparent, and Steerable Safety Moderation for AI Behavior
University of California, Berkeley(加州大学伯克利分校);University of Washington(华盛顿大学);Duke University(杜克大学);Carnegie Mellon University(卡内基梅隆大学);Stanford University(斯坦福大学)
Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training
Anthropic;Redwood Research(红木研究);Mila Quebec AI Institute(Mila魁北克人工智能研究所);University of Oxford(牛津大学);Alignment Research Center(对齐研究中心);Open Philanthropy(开放慈善);Apart Research
`Who would have thought of that!': A Hierarchical Topic Model for Extraction of Sarcasm-prevalent Topics and Sarcasm Detection
Indian Institute of Technology Bombay(印度理工学院孟买分校);Monash University(莫纳什大学);IITB-Monash Research Academy(印度理工学院孟买分校-莫纳什研究学院);IIT-BHU (Varanasi)(印度理工学院贝纳拉斯印度教大学分校(瓦拉纳西))
Think in Safety: Unveiling and Mitigating Safety Alignment Collapse in Multimodal Large Reasoning Model
Key Laboratory of Big Data & Artificial Intelligence in Transportation (Beijing Jiaotong University), Ministry of Education(大数据与人工智能交通联合实验室(北京交通大学));School of Computer Science and Technology, Beijing Jiaotong University(计算机科学与技术学院,北京交通大学);Tsinghua University(清华大学)