会员服务
AnchorKV: Safety-Aware KV Cache Compression via Soft Penalty with a Refusal Anchor
Arxiv
0+阅读 · 6月16日
Pulling The REINS: Training-Free Safety Alignment of Video Diffusion Models via Representation Steering
Arxiv
0+阅读 · 6月15日
Greed Is Learned: Visible Incentives as Reward-Hacking Triggers
Arxiv
0+阅读 · 6月15日
DOG-DPO:Dynamic Optimization in Geometry for Safety Alignment
Arxiv
0+阅读 · 6月14日
Cross-Generational Transfer of Adversarial Attacks Reveals Non-Monotonic Safety Alignment in LLMs
Arxiv
0+阅读 · 5月30日
SARSteer: Safeguarding Large Audio-Language Models via Safe-Ablated Refusal Steering
Arxiv
0+阅读 · 6月12日
SpeechJBB: Probing Safety Alignment and Comprehension in Large Audio Language Models under Code-Switched Speech
Arxiv
0+阅读 · 6月8日
FreoStream:Enhancing Stream Guardrails via Future-Aware Reasoning and Safety-Aligned Optimization
Arxiv
0+阅读 · 6月11日
When Autoregressive Consistency Hurts Safety Alignment
Arxiv
0+阅读 · 6月2日
Involuntary In-Context Learning: Exploiting Few-Shot Pattern Completion to Bypass Safety Alignment in GPT-5.4
Arxiv
0+阅读 · 6月3日
Involuntary In-Context Learning: Exploiting Few-Shot Pattern Completion to Bypass Safety Alignment in GPT-5.4
Arxiv
0+阅读 · 4月21日
Silenced Biases: The Dark Side LLMs Learned to Refuse
Arxiv
0+阅读 · 3月18日
SafeMath: Inference-time Safety improves Math Accuracy
Arxiv
0+阅读 · 3月26日
SecureBreak -- A dataset towards safe and secure models
Arxiv
0+阅读 · 3月23日
Self-Jailbreaking: Language Models Can Reason Themselves Out of Safety Alignment After Benign Reasoning Training
Arxiv
0+阅读 · 4月29日
参考链接
微信扫码咨询专知VIP会员
Top