AI Safety

HalluciDoctor: Mitigating Hallucinatory Toxicity in Visual Instruction Data CVPR 2024 Handling Long-Term Safety and Uncertainty in Safe Reinforcement Learning CoRL 2024 Hard-Label Based Small Query Black-Box Adversarial Attack WACV 2024 HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal ICML 2024 Harnessing Edge Information for Improved Robustness in Vision Transformers AAAI 2024 Here’s a Free Lunch: Sanitizing Backdoored Models with Model Merge ACL 2024 Holistic Automated Red Teaming for Large Language Models through Top-Down Test Case Generation and Multi-turn Interaction EMNLP 2024 Holistic Evaluation of Large Language Models: Assessing Robustness, Accuracy, and Toxicity for Real-World Applications NAACL 2024 HonestLLM: Toward an Honest and Helpful Large Language Model NeurIPS 2024 How Alignment and Jailbreak Work: Explain LLM Safety through Intermediate Hidden States EMNLP 2024 How does GPT-2 Predict Acronyms? Extracting and Understanding a Circuit via Mechanistic Interpretability AISTATS 2024 How Johnny Can Persuade LLMs to Jailbreak Them: Rethinking Persuasion to Challenge AI Safety by Humanizing LLMs ACL 2024 How Language Model Hallucinations Can Snowball ICML 2024 How should Conversational Agent systems respond to sexual harassment? EACL 2024 How Susceptible are Large Language Models to Ideological Manipulation? EMNLP 2024 How Susceptible Are LLMs to Logical Fallacies? COLING 2024 How to Catch an AI Liar: Lie Detection in Black-Box LLMs by Asking Unrelated Questions ICLR 2024 How Trustworthy are Open-Source LLMs? An Assessment under Malicious Demonstrations Shows their Vulnerabilities NAACL 2024 Hyper-opinion Evidential Deep Learning for Out-of-Distribution Detection NeurIPS 2024 IBD-PSC: Input-level Backdoor Detection via Parameter-oriented Scaling Consistency ICML 2024 Identifying the Risks of LM Agents with an LM-Emulated Sandbox ICLR 2024 IDGuard: Robust General Identity-centric POI Proactive Defense Against Face Editing Abuse CVPR 2024 I Don't Know: Explicit Modeling of Uncertainty with an [IDK] Token NeurIPS 2024 If You Want to Be Robust, Be Wary of Initialization NeurIPS 2024 Image Hijacks: Adversarial Images can Control Generative Models at Runtime ICML 2024