AI Safety

Layer-wise Swapping for Generalizable Multilingual Safety EACL 2026 Learning Multilingual Agentic Policy to Control Sycophancy EACL 2026 Learning Vision-Based Neural Network Controllers with Semi-Probabilistic Safety Guarantees AAAI 2026 LieCraft: A Multi-Agent Framework for Evaluating Deceptive Capabilities in Language Models AAAI 2026 Likelihood hacking in probabilistic program synthesis UAI 2026 Lipschitz Optimization for Formal Verification of Homographies CVPR 2026 LLaVAShield: Safeguarding Multimodal Multi-Turn Dialogues in Vision-Language Models CVPR 2026 LLM BiasScope: A Real-Time Bias Analysis Platform for Comparative LLM Evaluation EACL 2026 LLM Safety From Within: Detecting Harmful Content with Internal Representations ACL 2026 LLM Safety in Judicial AI: A Stress Test of Social Media Influence on Real-World Judgments AAAI 2026 LLM-VA: Resolving the Jailbreak-Overrefusal Trade-off via Vector Alignment ACL 2026 LoopLLM: Transferable Energy-Latency Attacks in LLMs via Repetitive Generation AAAI 2026 LUCID: Learning-Enabled Uncertainty-Aware Certification of Stochastic Dynamical Systems AAAI 2026 Lying with Truths: Open-Channel Multi-Agent Collusion for Belief Manipulation via Generative Montage ACL 2026 MAJIC: Markovian Adaptive Jailbreaking via Iterative Composition of Diverse Innovative Strategies AAAI 2026 Make Mechanistic Interpretability Auditable: A Call to Develop Guidelines via Continuous Collaborative Reviewing ACL 2026 MCA-Bench: A Multimodal Benchmark for Evaluating CAPTCHA Robustness Against VLM-based Attacks AAAI 2026 MCPTox: A Benchmark for Tool Poisoning on Real-World MCP Servers AAAI 2026 Medical Summarization in Practice: Design, Deployment, and Analysis of a Clinical Summarization System for a German Hospital EACL 2026 MedOmni-45°: A Safety–Performance Benchmark for Reasoning-Oriented LLMs in Medicine AAAI 2026 MedRiskEval: Medical Risk Evaluation Benchmark of Language Models, On the Importance of User Perspectives in Healthcare Settings EACL 2026 Mental Model-based Generation of Lies for Insider Threat Modeling AAAI 2026 MetaCipher: A Time-Persistent and Universal Multi-Agent Framework for Cipher-Based Jailbreak Attacks for LLMs AAAI 2026 MHB: Medical Hallucination Benchmark for Large Language Models in Complex Clinical Tasks AAAI 2026 MirrorShield: Towards Dynamic Adaptive Defense Against Jailbreaks via Entropy-Guided Mirror Crafting AAAI 2026