AI Safety

Unraveling LLM Jailbreaks Through Safety Knowledge Neurons EACL 2026 Unveiling AI Safety in Fine-tuning Quantized Model AAAI 2026 Unveiling the Attribute Misbinding Threat in Identity-Preserving Models AAAI 2026 Unveiling the Fragility of Vision-Language Models: Multi-Modal Adversarial Synergy via Texture-Constrained Perturbations and Cross-Modal Optimization AAAI 2026 Utterance-level Detection Framework for LLM-Involved Content Detection in Conversational Setting EACL 2026 Value-Aligned Prompt Moderation via Zero-Shot Agentic Rewriting for Safe Image Generation AAAI 2026 VeriFlow: Modeling Distributions for Neural Network Verification AAAI 2026 Verifying Neural Network Robustness with Dual Perturbations CVPR 2026 Verifying Nonlinear Neural Feedback Systems using Polyhedral Enclosures L4DC 2026 VILTA: A VLM-in-the-Loop Adversary for Enhancing Driving Policy Robustness AAAI 2026 VortexPIA: Indirect Prompt Injection Attack against LLMs for Efficient Extraction of User Privacy EACL 2026 Vulnerability-Aware Robust Multimodal Adversarial Training AAAI 2026 W2S-AlignTree: Weak-to-Strong Inference-Time Alignment for Large Language Models via Monte Carlo Tree Search AAAI 2026 WALKSAFE: Risk-aware Graph Random Walk with Bi-GRPO for LLM Safety AAAI 2026 When Can We Trust LLMs in Mental Health? Large-Scale Benchmarks for Reliable LLM Evaluation EACL 2026 When Do Language Models Endorse Limitations on Human Rights Principles? EACL 2026 When Human Preferences Flip: An Instance-Dependent Robust Loss for RLHF AAAI 2026 When Personalization Legitimizes Risks: Uncovering Safety Vulnerabilities in Personalized Dialogue Agents ACL 2026 When Prompt Optimization Becomes Jailbreaking: Adaptive Red-Teaming of Large Language Models EACL 2026 When Proxy Agents Disagree, Do Humans Mirror? Manipulating Human Behavior in Moral Dilemmas Through Agents AAAI 2026 When Safe Unimodal Inputs Collide: Optimizing Reasoning Chains for Cross-Modal Safety in Multimodal Large Language Models AAAI 2026 When Smiley Turns Hostile: Interpreting How Emojis Trigger LLMs’ Toxicity AAAI 2026 When the Model Said ‘No Comment’, We Knew Helpfulness Was Dead, Honesty Was Alive, and Safety Was Terrified EACL 2026 When Truth Is Overridden: Uncovering the Internal Origins of Sycophancy in Large Language Models AAAI 2026 When Understanding Becomes a Risk: Authenticity and Safety Risks in the Emerging Image Generation Paradigm CVPR 2026