牛哥精选 · 三个月

📋 全部 🤖 AI·大模型 ⚡ 效率工具 📝 深度技术 🚀 产品观察 💰 商业科技 🔓 开源项目 🎨 设计创意 📖 阅读推荐 🏷 资源合集 🌱 成长效率

🤖 AI·大模型 arXiv NLP 2026-06-09

MC-PDD: Masked Corpus-Level Pretraining Data Detection for Black-Box Large Language Models

提出黑盒场景下掩码语料级预训练数据检测新方法，助力LLM数据泄露评估

arXiv:2606.07996v1 Announce Type: new Abstract: Pretraining is fundamental to the development of Large Language Models (LLMs), yet the opacity of pret…

预训练数据检测黑盒llm 数据泄露掩码方法语料级检测

📝 深度技术 arXiv AI 2026-05-19

How to Train Your Advisor: Steering Black-Box LLMs with Advisor Models

用小型开源模型生成动态建议，巧妙引导黑盒大模型提升输出质量，无需修改权重。

arXiv:2510.02453v3 Announce Type: replace-cross Abstract: Frontier language models are deployed as black-box services, where model weights cannot be m…

advisor mo 黑盒llm 小模型训练动态建议模型引导

🤖 AI·大模型 arXiv AI 2026-05-19

FlipAttack: Jailbreak LLMs via Flipping

揭秘LLM从左到右理解的弱点：仅需在左侧加噪声，就能轻松绕过黑盒大模型的安全护栏。

arXiv:2410.02832v2 Announce Type: replace-cross Abstract: This paper proposes a simple yet effective jailbreak attack named FlipAttack against black-b…

flipattack 越狱攻击黑盒llm 安全漏洞左侧噪声

📅 日期

2026-05-20 2026-05-19

🐂 牛哥精选

MC-PDD: Masked Corpus-Level Pretraining Data Detection for Black-Box Large Language Models

How to Train Your Advisor: Steering Black-Box LLMs with Advisor Models

FlipAttack: Jailbreak LLMs via Flipping

📅 日期