AI Safety

DeDe: Detecting Backdoor Samples for SSL Encoders via Decoders CVPR 2025 Defending against Indirect Prompt Injection by Instruction Detection EMNLP 2025 Defending Against Repetitive Backdoor Attacks on Semi-Supervised Learning through Lens of Rate-Distortion-Perception Trade-Off WACV 2025 Defending Large Language Models against Jailbreak Attacks via Semantic Smoothing IJCNLP 2025 Defending LVLMs Against Vision Attacks Through Partial-Perception Supervision ICML 2025 Defense Against Prompt Injection Attack by Leveraging Attack Techniques ACL 2025 Defense against Prompt Injection Attacks via Mixture of Encodings NAACL 2025 Defensive Prompt Patch: A Robust and Generalizable Defense of Large Language Models against Jailbreak Attacks ACL 2025 Dehumanizing Machines: Mitigating Anthropomorphic Behaviors in Text Generation Systems ACL 2025 DELMAN: Dynamic Defense Against Large Language Model Jailbreaking with Model Editing ACL 2025 Democratic Training Against Universal Adversarial Perturbations ICLR 2025 DemonAgent: Dynamically Encrypted Multi-Backdoor Implantation Attack on LLM-based Agent EMNLP 2025 Demystify Verbosity Compensation Behavior of Large Language Models EMNLP 2025 Deontological Keyword Bias: The Impact of Modal Expressions on Normative Judgments of Language Models ACL 2025 DeTAM: Defending LLMs Against Jailbreak Attacks via Targeted Attention Modification ACL 2025 Detecting Backdoor Attacks in Federated Learning via Direction Alignment Inspection CVPR 2025 Detecting Child Objectification on Social Media: Challenges in Language Modeling ACL 2025 Detecting deepfakes and false ads through analysis of text and social engineering techniques COLING 2025 Detecting Stealthy Backdoor Samples based on Intra-class Distance for Large Language Models EMNLP 2025 Detecting Strategic Deception with Linear Probes ICML 2025 Deterministic Certification of Graph Neural Networks against Graph Poisoning Attacks with Arbitrary Perturbations CVPR 2025 Detoxifying Large Language Models via the Diversity of Toxic Samples EMNLP 2025 Detoxify-IT: An Italian Parallel Dataset for Text Detoxification ACL 2025 DF-MIA: A Distribution-Free Membership Inference Attack on Fine-Tuned Large Language Models AAAI 2025 DIESEL: A Lightweight Inference-Time Safety Enhancement for Language Models ACL 2025