AI Safety

From raw data to safety: Reducing conservatism by set expansion L4DC 2024 From Representational Harms to Quality-of-Service Harms: A Case Study on Llama 2 Safety Safeguards ACL 2024 From Shortcuts to Triggers: Backdoor Defense with Denoised PoE NAACL 2024 From Toxic to Trustworthy: Using Self-Distillation and Semi-supervised Methods to Refine Neural Networks AAAI 2024 Fundamental Limitations of Alignment in Large Language Models ICML 2024 Gaining Wisdom from Setbacks: Aligning Large Language Models via Mistake Analysis ICLR 2024 GaLileo: General Linear Relaxation Framework for Tightening Robustness Certification of Transformers AAAI 2024 Gameplay Filters: Robust Zero-Shot Safety through Adversarial Imagination CoRL 2024 Generalized constraint for probabilistic safe reinforcement learning L4DC 2024 Glue pizza and eat rocks - Exploiting Vulnerabilities in Retrieval-Augmented Generative Models EMNLP 2024 GNNCert: Deterministic Certification of Graph Neural Networks against Adversarial Perturbations ICLR 2024 GNNs Also Deserve Editing, and They Need It More Than Once ICML 2024 Goal Alignment: Re-analyzing Value Alignment Problems Using Human-Aware AI AAAI 2024 Goodhart's Law in Reinforcement Learning ICLR 2024 GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Cipher ICLR 2024 GPT-4 Jailbreaks Itself with Near-Perfect Success Using Self-Explanation EMNLP 2024 Gradient Alignment for Cross-Domain Face Anti-Spoofing CVPR 2024 Gradient-Based Language Model Red Teaming EACL 2024 Gradient Cuff: Detecting Jailbreak Attacks on Large Language Models by Exploring Refusal Loss Landscapes NeurIPS 2024 Gradient shaping for multi-constraint safe reinforcement learning L4DC 2024 GradSafe: Detecting Jailbreak Prompts for LLMs via Safety-Critical Gradient Analysis ACL 2024 Groningen Group E at SemEval-2024 Task 8: Detecting machine-generated texts through pre-trained language models augmented with explicit linguistic-stylistic features SemEval 2024 GuardBench: A Large-Scale Benchmark for Guardrail Models EMNLP 2024 GumbelSoft: Diversified Language Model Watermarking via the GumbelMax-trick ACL 2024 Hacking predictors means hacking cars: Using sensitivity analysis to identify trajectory prediction vulnerabilities for autonomous driving security L4DC 2024