https://guijiagi.com/tags/%E5%AE%89%E5%85%A8%E5%AF%B9%E9%BD%90/Recent content in 安全对齐 on 硅基 AGI · 智能体学习与测评Hugozh-cn本站内容采用 CC BY-NC-SA 4.0 国际许可协议授权Thu, 16 Jul 2026 10:05:00 +0800https://guijiagi.com/posts/b2-fd3b570e/Thu, 16 Jul 2026 10:05:00 +0800https://guijiagi.com/posts/b2-fd3b570e/全面梳理大模型安全对齐技术体系，涵盖RLHF、DPO、Constitutional AI及红队测试实践https://guijiagi.com/posts/article-63/Mon, 13 Jul 2026 03:20:00 +0800https://guijiagi.com/posts/article-63/系统梳理大模型安全对齐技术栈，从RLHF到Constitutional AI，从训练阶段到推理阶段的全链路安全方案https://guijiagi.com/posts/llm-jailbreak-landscape-dan-to-multi-turn/Tue, 30 Jun 2026 10:10:00 +0800https://guijiagi.com/posts/llm-jailbreak-landscape-dan-to-multi-turn/全面梳理2026年大模型越狱技术的演进路径，从早期DAN越狱到多轮诱导、多模态攻击，分析攻防博弈的深层逻辑