AI Safety

Backdoor Defense via Enhanced Splitting and Trap Isolation ICCV 2025 Backdoor Mitigation by Distance-Driven Detoxification ICCV 2025 Backdoor-Powered Prompt Injection Attacks Nullify Defense Methods EMNLP 2025 Backdoor Token Unlearning: Exposing and Defending Backdoors in Pretrained Language Models AAAI 2025 Back to Base: Towards Hands-Off Learning via Safe Resets with Reach-Avoid Safety Filters L4DC 2025 Backtracking Improves Generation Safety ICLR 2025 BadJudge: Backdoor Vulnerabilities of LLM-As-A-Judge ICLR 2025 BadRobot: Jailbreaking Embodied LLM Agents in the Physical World ICLR 2025 BadToken: Token-level Backdoor Attacks to Multi-modal Large Language Models CVPR 2025 BadVideo: Stealthy Backdoor Attack against Text-to-Video Generation ICCV 2025 BaxBench: Can LLMs Generate Correct and Secure Backends? ICML 2025 Benchmarking Large Language Models for Cryptanalysis and Side-Channel Vulnerabilities EMNLP 2025 Benchmarking Multi-National Value Alignment for Large Language Models ACL 2025 Beneath the Facade: Probing Safety Vulnerabilities in LLMs via Auto-Generated Jailbreak Prompts EMNLP 2025 Benign Samples Matter! Fine-tuning On Outlier Benign Samples Severely Breaks Safety ICML 2025 Better Private Distribution Testing by Leveraging Unverified Auxiliary Data COLT 2025 Better Red Teaming via Searching with Large Language Model ACL 2025 Beware of Your Po! Measuring and Mitigating AI Safety Risks in Role-Play Fine-Tuning of LLMs ACL 2025 Beyond Clean Training Data: A Versatile and Model-Agnostic Framework for Out-of-Distribution Detection with Contaminated Training Data CVPR 2025 Beyond Guardrails: Advanced Safety for Large Language Models — Monolingual, Multilingual and Multimodal Frontiers IJCNLP 2025 Beyond Hate Speech: NLP’s Challenges and Opportunities in Uncovering Dehumanizing Language EMNLP 2025 Beyond Human Judgment: A Bayesian Evaluation of LLMs’ Moral Values Understanding EMNLP 2025 Beyond Prompt Engineering: Robust Behavior Control in LLMs via Steering Target Atoms ACL 2025 Beyond Reactive Safety: Risk-Aware LLM Alignment via Long-Horizon Simulation ACL 2025 Beyond Single-Value Metrics: Evaluating and Enhancing LLM Unlearning with Cognitive Diagnosis ACL 2025