跨架构API提取成本控制利器，精准防护分数返回接口，有效降低数据泄露风险，安全高效。
arXiv:2608.01023v1 Announce Type: cross Abstract: We present Caliber, an output-perturbation defense against model extraction that formulates noise se…
从OpenAI与Hugging Face攻击事件切入，深度剖析AI对齐失败带来的存在性威胁
Article URL: https://www.lesswrong.com/posts/H6DDSEvrtCk8Sehfd/are-we-existentially-threatened-by-the-type-of-ai Comments URL: https://news.ycombinato…
新测试平台LACUNA精准评估大模型遗忘方法，定位精度是关键突破点。
arXiv:2607.02513v1 Announce Type: cross Abstract: LLMs memorize sensitive training data, including personally identifiable information (PII), creating…
研究如何通过战略性地扰动成对数据来操纵排名系统，揭示AI安全领域的新漏洞。
arXiv:2604.17805v2 Announce Type: replace-cross Abstract: Pairwise ranking systems based on Maximum Likelihood Estimation (MLE), such as the Bradley-T…
最新研究揭示大模型“遗忘”机制：事实显著性与微调如何双重影响模型知识删除的效果。
arXiv:2602.19612v5 Announce Type: replace Abstract: Machine Unlearning (MU) enables Large Language Models (LLMs) to remove unsafe or outdated informat…
提出公平微调方法，有效抵御机器学习模型的黑盒分布推断攻击，守护群体隐私。
arXiv:2606.01719v1 Announce Type: new Abstract: Machine learning models trained on sensitive data can inadvertently leak population-level information …
PRBench为概率鲁棒性评估提供了标准化基准，助你系统衡量AI模型的安全性。
arXiv:2511.01724v3 Announce Type: replace-cross Abstract: Deep learning models are notoriously vulnerable to imperceptible perturbations. Most existin…
研究表明，LLM在摘要任务中导出的向量表示并不中立，会泄露敏感信息，引发隐私安全新担忧。
arXiv:2605.26433v1 Announce Type: new Abstract: Large language model (LLM) summarization systems may pass compact vector representations of private in…
后量子时代多智能体AI治理系统，提出可证明安全的MAGIQ架构，解决新兴计算范式下的安全挑战。
arXiv:2605.06933v2 Announce Type: replace Abstract: Our computing ecosystem is being transformed by two emerging paradigms: the increased deployment o…
分布式学习中数据归因的脆弱性：单个参与者可操纵归因值大幅膨胀，挑战定价与审计可信度。
arXiv:2605.15520v1 Announce Type: cross Abstract: Data attribution has become an important component of pricing, auditing, and governance in machine l…