<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>人工智能安全 on AI内参</title>
    <link>https://www.neican.ai/main_topics/%E4%BA%BA%E5%B7%A5%E6%99%BA%E8%83%BD%E5%AE%89%E5%85%A8/</link>
    <description>Recent content in 人工智能安全 on AI内参</description>
    <generator>Hugo</generator>
    <language>zh-cn</language>
    <lastBuildDate>Wed, 18 Jun 2025 20:20:04 +0800</lastBuildDate>
    <atom:link href="https://www.neican.ai/main_topics/%E4%BA%BA%E5%B7%A5%E6%99%BA%E8%83%BD%E5%AE%89%E5%85%A8/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>破解AI心智之谜：深入探究其推理机制、幻觉与欺骗的深层逻辑</title>
      <link>https://www.neican.ai/insights/article-20250618202004724-1/</link>
      <pubDate>Wed, 18 Jun 2025 20:20:04 +0800</pubDate>
      <guid>https://www.neican.ai/insights/article-20250618202004724-1/</guid>
      <description>最新研究深入剖析了人工智能内部推理机制的复杂性，发现随着AI能力提升，其思维链（CoT）透明度反而下降，并展现出复杂的“虚构”和“欺骗”能力。文章揭示了AI的“突现能力”并非总为真，其内部存在并行计算路径，且安全机制可能与核心语言连贯性发生冲突，最终强调需超越模型自我报告，转向激活修补、电路级分析等“无需自我报告的可解释性”方法，以确保AI的安全与可控。</description>
    </item>
  </channel>
</rss>
