Safety

How Controllable Are Large Language Models? A Unified Evaluation across Behavioral Granularities ACL 2026 In-Context Representation Hijacking ACL 2026 Inertia in Moral and Value Judgments of Large Language Models ACL 2026 Interpretable Safety Alignment via SAE-Constructed Low-Rank Subspace Adaptation ACL 2026 IPS: In-Prompt Process Supervision for Short Video Content Moderation ACL 2026 JANUS: A Lightweight Framework for Jailbreaking Text-to-Image Models via Distribution Optimization CVPR 2026 JPU: Bridging Jailbreak Defense and Unlearning via On-Policy Path Rectification ACL 2026 Kernel-Based Safe Exploration in Deep Reinforcement Learning L4DC 2026 Knowing When Not to Answer: Lightweight KB-Aligned OOD Detection for Safe RAG ACL 2026 Knowledge Control for Responsible Generative AI: Bridging Academia, Industry, and Society ACL 2026 LAFaCT: Attribution-based Localization and Focused Sequential Analysis of Fact-Critical Tokens for Hallucination Detection ACL 2026 Learning to Conceal Risk: Controllable Multi-turn Red Teaming for LLMs in the Financial Domain ACL 2026 MirageBackdoor: A Stealthy Attack that Induces Think-Well-Answer-Wrong Reasoning ACL 2026 Mitigating Safety Context Amnesia in Multimodal Reasoning Models via Intent-Guided Safety Reasoning ACL 2026 More Thinking, Less Talking: Internalizing Deliberative Safety into LLM Parameters ACL 2026 Multimodal Safety Evaluation in Generative Agent Social Simulations ACL 2026 N-GLARE: An Non-Generative Latent Representation-Efficient LLM Safety Evaluator ACL 2026 OASIS: Mitigating Harmful Fine-tuning Attacks on LLMs via Orthogonal and Adaptive Safety Alignment Strategy ACL 2026 OS-Sentinel: Towards Safety-Enhanced Mobile GUI Agents via Hybrid Validation in Realistic Workflows ACL 2026 Persona-Grounded Safety Evaluation of AI Companions in Multi-Turn Conversations ACL 2026 Please refuse to answer me! Mitigating Over-Refusal in Large Language Models via Adaptive Contrastive Decoding ACL 2026 Probing the Safety Robustness of LLMs in Latent Space ACL 2026 Projecting Out the Malice: A Global Subspace Approach to LLM Detoxification ACL 2026 Provably Safe Offline-to-Online RL: Decoupling Learning from Data-Driven Safety Enforcement ACL 2026 Pruning Unsafe Tickets: A Resource-Efficient Framework for Safer and More Robust LLMs ACL 2026