<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>AI45研究成果</title>
    <link>https://ai45.shlab.org.cn/research/zh/</link>
    <description>Recent content on AI45研究成果</description>
    <image>
      <title>AI45研究成果</title>
      <url>https://ai45.shlab.org.cn/research/images/papermod-cover.png</url>
      <link>https://ai45.shlab.org.cn/research/images/papermod-cover.png</link>
    </image>
    <generator>Hugo -- 0.147.9</generator>
    <language>zh</language>
    <copyright>PaperMod Contributors</copyright>
    <lastBuildDate>Tue, 10 Feb 2026 00:00:00 +0000</lastBuildDate>
    <atom:link href="https://ai45.shlab.org.cn/research/zh/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>SafeVerse开源：构建安全可信的具身智能“孪生演练场”</title>
      <link>https://ai45.shlab.org.cn/research/zh/posts/safeverse%E5%85%AC%E4%BC%97%E5%8F%B7%E6%8E%A8%E9%80%81/</link>
      <pubDate>Tue, 10 Feb 2026 00:00:00 +0000</pubDate>
      <guid>https://ai45.shlab.org.cn/research/zh/posts/safeverse%E5%85%AC%E4%BC%97%E5%8F%B7%E6%8E%A8%E9%80%81/</guid>
      <description>一段普通视频分钟级生成可交互、符合物理规律的3D孪生演练场，支持攻防对抗与在线进化的开源平台。</description>
    </item>
    <item>
      <title>以循环为桥：循环 Transformers 能否弥合输出与表征的差距?</title>
      <link>https://ai45.shlab.org.cn/research/zh/posts/blog_test/</link>
      <pubDate>Fri, 16 Jan 2026 10:00:00 +0800</pubDate>
      <guid>https://ai45.shlab.org.cn/research/zh/posts/blog_test/</guid>
      <description>探索 Looped Transformer 如何通过内生循环机制，弥合大模型语言输出与内部潜意识表征之间的鸿沟。</description>
    </item>
    <item>
      <title>SafeWork-F1：前沿 AI 风险管理框架</title>
      <link>https://ai45.shlab.org.cn/research/zh/posts/safework-f1/</link>
      <pubDate>Fri, 25 Jul 2025 10:00:00 +0800</pubDate>
      <guid>https://ai45.shlab.org.cn/research/zh/posts/safework-f1/</guid>
      <description>推动“不易验证但不容试错”的远端灾难性风险，从“模糊的定义”到“精确的坐标”。</description>
    </item>
    <item>
      <title>SafeWork-T1：多模态大模型的安全推理训练加速器</title>
      <link>https://ai45.shlab.org.cn/research/zh/posts/safework-t1/</link>
      <pubDate>Mon, 21 Jul 2025 10:00:00 +0800</pubDate>
      <guid>https://ai45.shlab.org.cn/research/zh/posts/safework-t1/</guid>
      <description>面向多模态大模型打造的安全推理训练平台，专注于提升训练效率、保障推理可信，并支持跨领域快速迁移。</description>
    </item>
    <item>
      <title>潜伏在你手机里的智能特工</title>
      <link>https://ai45.shlab.org.cn/research/zh/posts/%E6%BD%9C%E4%BC%8F%E5%9C%A8%E4%BD%A0%E6%89%8B%E6%9C%BA%E9%87%8C%E7%9A%84%E6%99%BA%E8%83%BD%E7%89%B9%E5%B7%A5%E7%AB%AF%E4%BE%A7agent%E9%9C%80%E8%A6%81%E6%80%8E%E6%A0%B7%E7%9A%84%E5%AE%89%E5%85%A8%E5%B9%B4%E6%A3%80/</link>
      <pubDate>Fri, 18 Jul 2025 10:00:00 +0800</pubDate>
      <guid>https://ai45.shlab.org.cn/research/zh/posts/%E6%BD%9C%E4%BC%8F%E5%9C%A8%E4%BD%A0%E6%89%8B%E6%9C%BA%E9%87%8C%E7%9A%84%E6%99%BA%E8%83%BD%E7%89%B9%E5%B7%A5%E7%AB%AF%E4%BE%A7agent%E9%9C%80%E8%A6%81%E6%80%8E%E6%A0%B7%E7%9A%84%E5%AE%89%E5%85%A8%E5%B9%B4%E6%A3%80/</guid>
      <description>构建面向手机端侧AI助手的高保真、全方位安全评测框架，为手机成为用户的“第二大脑”构筑坚实的安全防线.</description>
    </item>
    <item>
      <title>SafeWork-E1: 永不疲倦的第二大脑-可控可信的清洁能源基座</title>
      <link>https://ai45.shlab.org.cn/research/zh/posts/safework_e1/</link>
      <pubDate>Mon, 14 Jul 2025 10:00:00 +0800</pubDate>
      <guid>https://ai45.shlab.org.cn/research/zh/posts/safework_e1/</guid>
      <description>针对清洁能源智能化转型中的安全可信挑战，实验室联合行业头部企业，共同打造该领域首个安全推理基座模型。</description>
    </item>
    <item>
      <title>SafeWork-R1：AI-45°平衡律引领下的智能与安全协同进化</title>
      <link>https://ai45.shlab.org.cn/research/zh/posts/safework-r1/</link>
      <pubDate>Sat, 12 Jul 2025 09:00:00 +0800</pubDate>
      <guid>https://ai45.shlab.org.cn/research/zh/posts/safework-r1/</guid>
      <description>一个兼具卓越安全性与通用能力的多模态推理模型，具备更具表现力的语言、更可靠的逻辑推理能力、更友好的人机交互行为。该模型基于我们自研的安全加固框架SafeLadder构建而成。</description>
    </item>
    <item>
      <title>SafeWork-V1：自动形式化验证系统</title>
      <link>https://ai45.shlab.org.cn/research/zh/posts/safework-v1/</link>
      <pubDate>Sat, 12 Jul 2025 09:00:00 +0800</pubDate>
      <guid>https://ai45.shlab.org.cn/research/zh/posts/safework-v1/</guid>
      <description>自动形式化验证系统，构建起数学上可证明的人工智能安全保障系统。</description>
    </item>
  </channel>
</rss>
