SOTAVerified

Safety Alignment

Papers

Recently AddedMost HypedMost ActiveNeeds VerificationMost Verified

Showing 151–200 of 288 papers

TitleStatusHype
Mimicking User Data: On Mitigating Fine-Tuning Risks in Closed Large Language Models—0
Model Card and Evaluations for Claude Models—0
Model-Editing-Based Jailbreak against Safety-aligned Large Language Models—0
Model Merging and Safety Alignment: One Bad Model Spoils the Bunch—0
More is Less: The Pitfalls of Multi-Model Synthetic Preference Data in DPO Safety Alignment—0
Multilingual Blending: LLM Safety Alignment Evaluation with Language Mixture—0
Na'vi or Knave: Jailbreaking Language Models via Metaphorical Avatars—0
NeuRel-Attack: Neuron Relearning for Safety Disalignment in Large Language Models—0
No Free Lunch for Defending Against Prefilling Attack by In-Context Learning—0
Noise Injection Systemically Degrades Large Language Model Safety Guardrails—0
No Two Devils Alike: Unveiling Distinct Mechanisms of Fine-tuning Attacks—0
Off-Policy Risk Assessment in Markov Decision Processes—0
One Trigger Token Is Enough: A Defense Strategy for Balancing Safety and Usability in Large Language Models—0
RLHFPoison: Reward Poisoning Attack for Reinforcement Learning with Human Feedback in Large Language Models—0
On the Intrinsic Self-Correction Capability of LLMs: Uncertainty and Latent Concept—0
PathSeeker: Exploring LLM Security Vulnerabilities with a Reinforcement Learning-Based Jailbreak Approach—0
PEFT-as-an-Attack! Jailbreaking Language Models during Federated Parameter-Efficient Fine-Tuning—0
PKU-SafeRLHF: Towards Multi-Level Safety Alignment for LLMs with Human Preference—0
Playing Language Game with LLMs Leads to Jailbreaking—0
PoisonSwarm: Universal Harmful Information Synthesis via Model Crowdsourcing—0
Safety is Not Only About Refusal: Reasoning-Enhanced Fine-tuning for Interpretable LLM Safety—0
SafeVid: Toward Safety Aligned Video Large Multimodal Models—0
SafeVLA: Towards Safety Alignment of Vision-Language-Action Model via Constrained Learning—0
SAFEx: Analyzing Vulnerabilities of MoE-Based LLMs via Stable Safety-critical Expert Identification—0
SAGE-RT: Synthetic Alignment data Generation for Safety Evaluation and Red Teaming—0
SaLoRA: Safety-Alignment Preserved Low-Rank Adaptation—0
SciSafeEval: A Comprehensive Benchmark for Safety Alignment of Large Language Models in Scientific Tasks—0
Security Assessment of DeepSeek and GPT Series Models against Jailbreak Attacks—0
SecurityLingua: Efficient Defense of LLM Jailbreak Attacks via Security-Aware Prompt Compression—0
Seeing the Threat: Vulnerabilities in Vision-Language Models to Adversarial Attack—0
SelfDefend: LLMs Can Defend Themselves against Jailbreaking in a Practical Manner—0
Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models—0
Shape it Up! Restoring LLM Safety during Finetuning—0
Smaller Large Language Models Can Do Moral Self-Correction—0
SOSBENCH: Benchmarking Safety Alignment on Scientific Knowledge—0
SPA-VL: A Comprehensive Safety Preference Alignment Dataset for Vision Language Models—0
SPIN: Self-Supervised Prompt INjection—0
STAR-1: Safer Alignment of Reasoning LLMs with 1K Data—0
sudoLLM : On Multi-role Alignment of Language Models—0
Superficial Safety Alignment Hypothesis—0
Survey of Vulnerabilities in Large Language Models Revealed by Adversarial Attacks—0
The dark deep side of DeepSeek: Fine-tuning attacks against the safety alignment of CoT-enabled models—0
The Geometry of Refusal in Large Language Models: Concept Cones and Representational Independence—0
The Multilingual Alignment Prism: Aligning Global and Local Preferences to Reduce Harm—0
Thought Manipulation: External Thought Can Be Efficient for Large Reasoning Models—0
Towards Comprehensive Post Safety Alignment of Large Language Models via Safety Patching—0
Towards Inference-time Category-wise Safety Steering for Large Language Models—0
Towards NSFW-Free Text-to-Image Generation via Safety-Constraint Direct Preference Optimization—0
Towards Safe AI Clinicians: A Comprehensive Study on Large Language Model Jailbreaking in Healthcare—0
Towards Understanding the Fragility of Multilingual LLMs against Fine-Tuning Attacks—0
Show:102550
← PrevPage 4 of 6Next →

No leaderboard results yet.