Safety

Accommodation and Epistemic Vigilance: A Pragmatic Account of Why LLMs Fail to Challenge Harmful Beliefs ACL 2026 A Functionality-Grounded Benchmark for Evaluating Web Agents in E-commerce Domains ACL 2026 A Lightweight Explainable Guardrail for Prompt Safety ACL 2026 AlignCultura: Towards Culturally Aligned Large Language Models? ACL 2026 Answering the Wrong Question: Reasoning Trace Inversion for Abstention in LLMs ACL 2026 ARES: Adaptive Red-Teaming and End-to-End Repair of Policy-Reward System ACL 2026 ARGUS: Defending Against Multimodal Indirect Prompt Injection via Steering Instruction-Following Behavior CVPR 2026 Backdoors in RLVR: Jailbreak Backdoors in LLMs From Verifiable Reward ACL 2026 Benchmarking Web Agent Safety under E-commerce Deceptive Interfaces ACL 2026 Between a Rock and a Hard Place: The Tension Between Ethical Reasoning and Safety Alignment in LLMs ACL 2026 Beyond Surface-Level Detection: Towards Cognitive-Driven Defense Against Jailbreak Attacks via Meta-Operations Reasoning ACL 2026 Beyond Text Prompts: Precise Concept Erasure through Text-Image Collaboration CVPR 2026 CachePrune: Teaching LLMs What Not to Follow via KV-Cache Editing ACL 2026 Calibrating Inference Time Alignment with Sequence-level Risk Accumulation ACL 2026 Can Factual Opinions Be Edited (Manipulated) in Large Language Models? ACL 2026 Can LLM Safety Be Ensured by Constraining Parameter Regions? ACL 2026 CAP: Controllable Alignment Prompting for Unlearning in LLMs ACL 2026 COMPASS: A Framework for Evaluating Organization-Specific Policy Alignment in LLMs ACL 2026 Confident, Calibrated, or Complicit: Safety Alignment and Ideological Bias in LLM Hate Speech Detection ACL 2026 Constraint-Aware Reinforcement Learning via Adaptive Action Scaling L4DC 2026 ContextLens: Modeling Imperfect Privacy and Safety Context for Legal Compliance ACL 2026 CRISP: Persistent Concept Unlearning via Sparse Autoencoders ACL 2026 Critical-CoT: A Robust Defense Framework against Reasoning-Level Backdoor Attacks in Large Language Models ACL 2026 CrossGuard: Safeguarding MLLMs against Joint-Modal Implicit Malicious Attacks ACL 2026 CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning ACL 2026