<!DOCTYPE html>
<html lang="zh-CN">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>AI Rationality Research - CMU Study</title>
<link rel="preconnect" href="https://fonts.googleapis.com">
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
<link href="https://fonts.googleapis.com/css2?family=Noto+Serif+SC:wght@400;500;600;700&family=Playfair+Display:ital,wght@0,400;0,500;0,600;0,700;1,400;1,500&family=Inter:wght@300;400;500;600&display=swap" rel="stylesheet">
<style>
* {
margin: 0;
padding: 0;
box-sizing: border-box;
}
:root {
--background: #f5f2ed;
--foreground: #3d3833;
--card: #faf8f5;
--muted-foreground: #6b655d;
--morandi-pink: #b8908a;
--morandi-blue: #7a9aaa;
--morandi-green: #8ba896;
--morandi-purple: #9b8fa8;
--morandi-rose: #c49a9a;
--morandi-taupe: #8b8075;
--morandi-beige: #e8e2d9;
--morandi-cream: #f9f6f1;
--border: #ddd8d0;
}
html {
scroll-behavior: smooth;
}
body {
font-family: 'Inter', 'Noto Serif SC', sans-serif;
background: linear-gradient(180deg, var(--morandi-cream) 0%, var(--morandi-beige) 50%, #ebe7e0 100%);
color: var(--foreground);
line-height: 1.6;
min-height: 100vh;
position: relative;
}
body::before {
content: '';
position: fixed;
inset: 0;
background-image: url("data:image/svg+xml,%3Csvg viewBox='0 0 256 256' xmlns='http://www.w3.org/2000/svg'%3E%3Cfilter id='noise'%3E%3CfeTurbulence type='fractalNoise' baseFrequency='0.8' numOctaves='4' stitchTiles='stitch'/%3E%3C/filter%3E%3Crect width='100%25' height='100%25' filter='url(%23noise)'/%3E%3C/svg%3E");
opacity: 0.03;
pointer-events: none;
mix-blend-mode: multiply;
z-index: 1;
}
.container {
max-width: 1200px;
margin: 0 auto;
padding: 0 24px;
position: relative;
z-index: 2;
}
/* Header Section */
header {
padding: 80px 0 60px;
text-align: center;
}
.label {
display: flex;
align-items: center;
justify-content: center;
gap: 12px;
margin-bottom: 40px;
}
.label-line {
width: 48px;
height: 1px;
background: var(--morandi-taupe);
}
.label-text {
font-size: 11px;
letter-spacing: 0.3em;
text-transform: uppercase;
color: var(--morandi-taupe);
font-weight: 500;
}
h1 {
font-family: 'Playfair Display', serif;
font-size: clamp(2.5rem, 6vw, 4.5rem);
font-weight: 600;
line-height: 1.15;
margin-bottom: 24px;
color: var(--foreground);
}
h1 .italic-pink {
font-style: italic;
color: var(--morandi-pink);
}
h1 .italic-blue {
font-style: italic;
color: var(--morandi-blue);
}
h1 .taupe {
color: var(--morandi-taupe);
}
.subtitle {
font-family: 'Noto Serif SC', serif;
font-size: clamp(1rem, 2vw, 1.25rem);
color: var(--muted-foreground);
max-width: 600px;
margin: 0 auto 40px;
}
.intro-cards {
display: grid;
grid-template-columns: repeat(auto-fit, minmax(280px, 1fr));
gap: 20px;
max-width: 900px;
margin: 0 auto;
}
.card {
background: var(--card);
border-radius: 12px;
padding: 24px;
box-shadow:
0 4px 6px -1px rgba(89, 79, 71, 0.08),
0 2px 4px -2px rgba(89, 79, 71, 0.06),
0 0 0 1px rgba(89, 79, 71, 0.04);
border: 1px solid var(--border);
transition: all 0.3s ease;
}
.card:hover {
box-shadow:
0 10px 15px -3px rgba(89, 79, 71, 0.12),
0 4px 6px -4px rgba(89, 79, 71, 0.08),
0 0 0 1px rgba(89, 79, 71, 0.06);
transform: translateY(-2px);
}
.card-flex {
display: flex;
align-items: flex-start;
gap: 16px;
}
.icon-circle {
width: 44px;
height: 44px;
border-radius: 50%;
display: flex;
align-items: center;
justify-content: center;
flex-shrink: 0;
}
.icon-circle.pink {
background: rgba(184, 144, 138, 0.15);
}
.icon-circle.blue {
background: rgba(122, 154, 170, 0.15);
}
.icon-circle svg {
width: 22px;
height: 22px;
}
.icon-circle.pink svg {
color: var(--morandi-pink);
}
.icon-circle.blue svg {
color: var(--morandi-blue);
}
.card h3 {
font-family: 'Noto Serif SC', serif;
font-size: 1.125rem;
font-weight: 600;
margin-bottom: 8px;
color: var(--foreground);
}
.card p {
font-size: 0.875rem;
color: var(--muted-foreground);
line-height: 1.6;
}
/* Section Styles */
section {
padding: 60px 0;
}
.section-header {
text-align: center;
margin-bottom: 48px;
}
.decorative-line {
width: 96px;
height: 1px;
background: linear-gradient(90deg, transparent 0%, var(--morandi-taupe) 20%, var(--morandi-taupe) 80%, transparent 100%);
margin: 0 auto 24px;
}
.section-header h2 {
font-family: 'Playfair Display', serif;
font-size: clamp(1.75rem, 4vw, 2.5rem);
font-weight: 600;
margin-bottom: 8px;
color: var(--foreground);
}
.section-header .en {
font-size: 0.875rem;
color: var(--muted-foreground);
}
/* Core Finding */
.core-finding {
background: linear-gradient(135deg, rgba(184, 144, 138, 0.08) 0%, rgba(122, 154, 170, 0.08) 100%);
border-radius: 20px;
padding: 48px;
border: 1px solid var(--border);
display: flex;
align-items: center;
gap: 32px;
}
.core-icon {
width: 80px;
height: 80px;
border-radius: 50%;
background: rgba(184, 144, 138, 0.2);
display: flex;
align-items: center;
justify-content: center;
flex-shrink: 0;
}
.core-icon svg {
width: 40px;
height: 40px;
color: var(--morandi-pink);
}
.core-content p {
font-family: 'Noto Serif SC', serif;
font-size: clamp(1.125rem, 2.5vw, 1.5rem);
line-height: 1.7;
margin-bottom: 16px;
}
.core-content .highlight-blue {
color: var(--morandi-blue);
font-weight: 600;
}
.core-content .highlight-pink {
color: var(--morandi-pink);
font-weight: 600;
}
.core-content .en-text {
font-family: 'Inter', sans-serif;
font-size: 0.9375rem;
color: var(--muted-foreground);
font-style: italic;
}
/* Four Cards Grid */
.four-grid {
display: grid;
grid-template-columns: repeat(auto-fit, minmax(320px, 1fr));
gap: 24px;
}
.discovery-card {
background: var(--card);
border-radius: 16px;
padding: 32px;
box-shadow:
0 4px 6px -1px rgba(89, 79, 71, 0.08),
0 2px 4px -2px rgba(89, 79, 71, 0.06),
0 0 0 1px rgba(89, 79, 71, 0.04);
border: 1px solid var(--border);
transition: all 0.3s ease;
}
.discovery-card:hover {
box-shadow:
0 10px 15px -3px rgba(89, 79, 71, 0.12),
0 4px 6px -4px rgba(89, 79, 71, 0.08),
0 0 0 1px rgba(89, 79, 71, 0.06);
}
.discovery-header {
display: flex;
align-items: center;
gap: 16px;
margin-bottom: 20px;
}
.discovery-icon {
width: 56px;
height: 56px;
border-radius: 12px;
display: flex;
align-items: center;
justify-content: center;
transition: transform 0.3s ease;
}
.discovery-card:hover .discovery-icon {
transform: scale(1.1);
}
.discovery-icon.rose {
background: rgba(196, 154, 154, 0.15);
}
.discovery-icon.blue {
background: rgba(122, 154, 170, 0.15);
}
.discovery-icon.purple {
background: rgba(155, 143, 168, 0.15);
}
.discovery-icon.green {
background: rgba(139, 168, 150, 0.15);
}
.discovery-icon svg {
width: 28px;
height: 28px;
}
.discovery-icon.rose svg {
color: var(--morandi-rose);
}
.discovery-icon.blue svg {
color: var(--morandi-blue);
}
.discovery-icon.purple svg {
color: var(--morandi-purple);
}
.discovery-icon.green svg {
color: var(--morandi-green);
}
.discovery-title {
flex: 1;
}
.discovery-title .num {
font-size: 11px;
letter-spacing: 0.1em;
text-transform: uppercase;
color: var(--muted-foreground);
display: block;
margin-bottom: 4px;
}
.discovery-title h3 {
font-family: 'Noto Serif SC', serif;
font-size: 1.25rem;
font-weight: 600;
color: var(--foreground);
}
.discovery-card .en-title {
font-weight: 600;
color: var(--foreground);
margin-bottom: 12px;
font-size: 0.9375rem;
}
.discovery-card .desc {
font-size: 0.875rem;
color: var(--muted-foreground);
line-height: 1.7;
margin-bottom: 16px;
}
.discovery-card .quote {
font-size: 0.8125rem;
color: var(--morandi-taupe);
font-style: italic;
line-height: 1.7;
padding-left: 16px;
border-left: 2px solid;
}
.discovery-card .quote.rose-border {
border-color: rgba(196, 154, 154, 0.3);
}
.discovery-card .quote.blue-border {
border-color: rgba(122, 154, 170, 0.3);
}
.discovery-card .quote.purple-border {
border-color: rgba(155, 143, 168, 0.3);
}
.discovery-card .quote.green-border {
border-color: rgba(139, 168, 150, 0.3);
}
/* Quote Section */
.quote-section {
position: relative;
padding: 40px 0;
}
.quote-box {
background: linear-gradient(135deg, var(--morandi-beige) 0%, var(--card) 100%);
border-radius: 20px;
padding: 60px 48px 48px;
border: 1px solid var(--border);
text-align: center;
position: relative;
}
.quote-icon {
position: absolute;
top: -24px;
left: 50%;
transform: translateX(-50%);
width: 48px;
height: 48px;
border-radius: 50%;
background: rgba(184, 144, 138, 0.15);
display: flex;
align-items: center;
justify-content: center;
}
.quote-icon svg {
width: 24px;
height: 24px;
color: var(--morandi-pink);
}
.quote-box p {
font-family: 'Noto Serif SC', serif;
font-size: clamp(1.125rem, 2.5vw, 1.5rem);
line-height: 1.8;
margin-bottom: 16px;
}
.quote-box .highlight {
color: var(--morandi-pink);
font-weight: 600;
}
.quote-box .en {
font-family: 'Inter', sans-serif;
font-size: 0.875rem;
color: var(--muted-foreground);
font-style: italic;
}
/* Stats Section */
.stats-grid {
display: grid;
grid-template-columns: repeat(auto-fit, minmax(180px, 1fr));
gap: 24px;
}
.stat-card {
background: var(--card);
border-radius: 12px;
padding: 28px;
box-shadow:
0 4px 6px -1px rgba(89, 79, 71, 0.08),
0 2px 4px -2px rgba(89, 79, 71, 0.06),
0 0 0 1px rgba(89, 79, 71, 0.04);
border: 1px solid var(--border);
text-align: center;
}
.stat-card .number {
font-family: 'Playfair Display', serif;
font-size: 2rem;
font-weight: 600;
margin-bottom: 8px;
}
.stat-card.pink-stat .number {
color: var(--morandi-pink);
}
.stat-card.blue-stat .number {
color: var(--morandi-blue);
}
.stat-card.green-stat .number {
color: var(--morandi-green);
}
.stat-card .label-cn {
font-size: 0.875rem;
color: var(--muted-foreground);
margin-bottom: 4px;
}
.stat-card .label-en {
font-size: 0.75rem;
color: var(--morandi-taupe);
}
/* Footer */
footer {
padding: 60px 0 40px;
}
.footer-line {
width: 100%;
height: 1px;
background: linear-gradient(90deg, transparent 0%, var(--morandi-taupe) 20%, var(--morandi-taupe) 80%, transparent 100%);
margin-bottom: 40px;
}
.footer-content {
display: flex;
flex-direction: column;
align-items: center;
gap: 24px;
text-align: center;
}
<span class="mention-invalid">@media</span> (min-width: 768px) {
.footer-content {
flex-direction: row;
justify-content: space-between;
text-align: left;
}
}
.footer-left h3 {
font-family: 'Noto Serif SC', serif;
font-size: 1.25rem;
font-weight: 600;
margin-bottom: 4px;
}
.footer-left p {
font-size: 0.875rem;
color: var(--muted-foreground);
}
.footer-right {
display: flex;
align-items: center;
gap: 8px;
font-size: 0.875rem;
color: var(--muted-foreground);
}
.footer-right .arrow {
color: var(--morandi-taupe);
}
.footer-right .arxiv {
color: var(--morandi-taupe);
}
.footer-author {
text-align: center;
margin-top: 40px;
font-size: 0.75rem;
color: var(--muted-foreground);
}
/* Responsive */
<span class="mention-invalid">@media</span> (max-width: 768px) {
header {
padding: 60px 0 40px;
}
.core-finding {
flex-direction: column;
text-align: center;
padding: 32px;
}
.core-icon {
width: 64px;
height: 64px;
}
.core-icon svg {
width: 32px;
height: 32px;
}
section {
padding: 40px 0;
}
.quote-box {
padding: 48px 24px 36px;
}
}
/* Animations */
<span class="mention-invalid">@keyframes</span> fadeInUp {
from {
opacity: 0;
transform: translateY(30px);
}
to {
opacity: 1;
transform: translateY(0);
}
}
.animate-on-scroll {
opacity: 0;
animation: fadeInUp 0.8s ease forwards;
}
.delay-1 { animation-delay: 0.1s; }
.delay-2 { animation-delay: 0.2s; }
.delay-3 { animation-delay: 0.3s; }
.delay-4 { animation-delay: 0.4s; }
.delay-5 { animation-delay: 0.5s; }
</style>
<base target="_blank">
</head>
<body>
<div class="container">
<!-- Header Section -->
<header>
<div class="label animate-on-scroll">
<div class="label-line"></div>
<span class="label-text">CMU Research · AI Safety</span>
<div class="label-line"></div>
</div>
<h1 class="animate-on-scroll delay-1">
<span class="italic-pink">Rational</span> Synthesizers
<br>
<span class="taupe">or</span> <span class="italic-blue">Heuristic</span> Followers?
</h1>
<p class="subtitle animate-on-scroll delay-2">
卡内基梅隆大学最新研究揭开AI决策黑箱的裂痕
</p>
<div class="intro-cards">
<div class="card animate-on-scroll delay-3">
<div class="card-flex">
<div class="icon-circle pink">
<svg xmlns="http://www.w3.org/2000/svg" fill="none" viewBox="0 0 24 24" stroke="currentColor" stroke-width="1.5">
<path stroke-linecap="round" stroke-linejoin="round" d="M9.813 15.904L9 18.75l-.813-2.846a4.5 4.5 0 00-3.09-3.09L2.25 12l2.846-.813a4.5 4.5 0 003.09-3.09L9 5.25l.813 2.846a4.5 4.5 0 003.09 3.09L15.75 12l-2.846.813a4.5 4.5 0 00-3.09 3.09zM18.259 8.715L18 9.75l-.259-1.035a3.375 3.375 0 00-2.455-2.456L14.25 6l1.036-.259a3.375 3.375 0 002.455-2.456L18 2.25l.259 1.035a3.375 3.375 0 002.456 2.456L21.75 6l-1.035.259a3.375 3.375 0 00-2.456 2.456zM16.894 20.567L16.5 21.75l-.394-1.183a2.25 2.25 0 00-1.423-1.423L13.5 18.75l1.183-.394a2.25 2.25 0 001.423-1.423l.394-1.183.394 1.183a2.25 2.25 0 001.423 1.423l1.183.394-1.183.394a2.25 2.25 0 00-1.423 1.423z" />
</svg>
</div>
<div>
<h3>AI 真的理性吗?</h3>
<p>还是它只是一个读过很多书但耳根子很软的"复读机"?</p>
</div>
</div>
</div>
<div class="card animate-on-scroll delay-4">
<div class="card-flex">
<div class="icon-circle blue">
<svg xmlns="http://www.w3.org/2000/svg" fill="none" viewBox="0 0 24 24" stroke="currentColor" stroke-width="1.5">
<path stroke-linecap="round" stroke-linejoin="round" d="M9.75 3.104v5.714a2.25 2.25 0 01-.659 1.591L5 14.5M9.75 3.104c-.251.023-.501.05-.75.082m.75-.082a24.301 24.301 0 014.5 0m0 0v5.714c0 .597.237 1.17.659 1.591L19.8 15.3M14.25 3.104c.251.023.501.05.75.082M19.8 15.3l-1.57.393A9.065 9.065 0 0112 15a9.065 9.065 0 00-6.23-.693L5 14.5m14.8.8l1.402 1.402c1.232 1.232.65 3.318-1.067 3.611A48.309 48.309 0 0112 21c-2.773 0-5.491-.235-8.135-.687-1.718-.293-2.3-2.379-1.067-3.61L5 14.5" />
</svg>
</div>
<div>
<h3>深度解读论文</h3>
<p>Analyzing LLMs in RAG-based Question-Answering</p>
</div>
</div>
</div>
</div>
</header>
<!-- Core Finding Section -->
<section>
<div class="section-header animate-on-scroll">
<div class="decorative-line"></div>
<h2>核心发现</h2>
<p class="en">Core Findings</p>
</div>
<div class="core-finding animate-on-scroll delay-2">
<div class="core-icon">
<svg xmlns="http://www.w3.org/2000/svg" fill="none" viewBox="0 0 24 24" stroke="currentColor" stroke-width="1.5">
<path stroke-linecap="round" stroke-linejoin="round" d="M12 9v3.75m-9.303 3.376c-.866 1.5.217 3.374 1.948 3.374h14.71c1.73 0 2.813-1.874 1.948-3.374L13.949 3.378c-.866-1.5-3.032-1.5-3.898 0L2.697 16.126zM12 15.75h.007v.008H12v-.008z" />
</svg>
</div>
<div class="core-content">
<p>
我们以为 AI 是<span class="highlight-blue">客观的法官</span>,
但数据表明,它们更像是固执的<span class="highlight-pink">"经验主义者"</span>:
</p>
<p class="en-text">
We assume AI acts as an objective judge, but data shows they are stubborn "heuristic followers":
the larger the model, the less plastic it becomes, and they are easily fooled by repetitive information.
</p>
</div>
</div>
</section>
<!-- Four Key Discoveries -->
<section>
<div class="section-header animate-on-scroll">
<div class="decorative-line"></div>
<h2>四大关键发现</h2>
<p class="en">Four Key Discoveries</p>
</div>
<div class="four-grid">
<!-- Card 1 -->
<div class="discovery-card animate-on-scroll delay-1">
<div class="discovery-header">
<div class="discovery-icon rose">
<svg xmlns="http://www.w3.org/2000/svg" fill="none" viewBox="0 0 24 24" stroke="currentColor" stroke-width="1.5">
<path stroke-linecap="round" stroke-linejoin="round" d="M2.25 18L9 11.25l4.306 4.307a11.95 11.95 0 015.814-5.519l2.74-1.22m0 0l-5.94-2.28m5.94 2.28l-2.28 5.941" />
</svg>
</div>
<div class="discovery-title">
<span class="num">01</span>
<h3>经验法则追随者</h3>
</div>
</div>
<p class="en-title">Heuristic Followers</p>
<p class="desc">
我们打破了 AI 是"理性综合者"的幻想。研究发现,AI 在处理冲突信息时,往往不进行深度逻辑分析,而是依赖简单的统计捷径。
</p>
<p class="quote rose-border">
We shatter the illusion of AI as a "Rational Synthesizer." Research shows that when facing conflicting info, AI often skips deep logical analysis and relies on simple statistical shortcuts.
</p>
</div>
<!-- Card 2 -->
<div class="discovery-card animate-on-scroll delay-2">
<div class="discovery-header">
<div class="discovery-icon blue">
<svg xmlns="http://www.w3.org/2000/svg" fill="none" viewBox="0 0 24 24" stroke="currentColor" stroke-width="1.5">
<path stroke-linecap="round" stroke-linejoin="round" d="M12 9v3.75m0-10.036A11.959 11.959 0 013.598 6 11.99 11.99 0 003 9.75c0 5.592 3.824 10.29 9 11.622 5.176-1.332 9-6.03 9-11.622 0-1.31-.21-2.57-.598-3.75h-.152c-3.196 0-6.1-1.249-8.25-3.286zm0 13.036h.008v.008H12v-.008z" />
</svg>
</div>
<div class="discovery-title">
<span class="num">02</span>
<h3>可塑性悖论</h3>
</div>
</div>
<p class="en-title">The Plasticity Paradox</p>
<p class="desc">
一个反直觉的发现——模型参数规模越大(如 Llama-3 70B),反而越难以接受新的事实证据,表现出极强的"知识惯性"。
</p>
<p class="quote blue-border">
A counter-intuitive finding: the larger the model scale, the harder it is for it to accept new factual evidence, showing strong "knowledge inertia."
</p>
</div>
<!-- Card 3 -->
<div class="discovery-card animate-on-scroll delay-3">
<div class="discovery-header">
<div class="discovery-icon purple">
<svg xmlns="http://www.w3.org/2000/svg" fill="none" viewBox="0 0 24 24" stroke="currentColor" stroke-width="1.5">
<path stroke-linecap="round" stroke-linejoin="round" d="M19.5 12c0-1.232-.046-2.453-.138-3.662a4.006 4.006 0 00-3.7-3.7 48.678 48.678 0 00-7.324 0 4.006 4.006 0 00-3.7 3.7c-.017.22-.032.441-.046.662M19.5 12l3-3m-3 3l-3-3m-12 3c0 1.232.046 2.453.138 3.662a4.006 4.006 0 003.7 3.7 48.656 48.656 0 007.324 0 4.006 4.006 0 003.7-3.7c.017-.22.032-.441.046-.662M4.5 12l3 3m-3-3l-3 3" />
</svg>
</div>
<div class="discovery-title">
<span class="num">03</span>
<h3>虚幻真相效应</h3>
</div>
</div>
<p class="en-title">Illusory Truth Effect</p>
<p class="desc">
AI 竟然会被"车轱辘话"忽悠?实验证明,简单重复的冗余信息,比高质量的独立证据更能左右 AI 的判断。这意味着真相可以被"制造"。
</p>
<p class="quote purple-border">
Can AI be fooled by repetition? Experiments prove that simply repeating the same point persuades AI more than high-quality, independent evidence. Truth can be "manufactured."
</p>
</div>
<!-- Card 4 -->
<div class="discovery-card animate-on-scroll delay-4">
<div class="discovery-header">
<div class="discovery-icon green">
<svg xmlns="http://www.w3.org/2000/svg" fill="none" viewBox="0 0 24 24" stroke="currentColor" stroke-width="1.5">
<path stroke-linecap="round" stroke-linejoin="round" d="M7.5 8.25h9m-9 3H12m-9.75 1.51c0 1.6 1.123 2.994 2.707 3.227 1.129.166 2.27.293 3.423.379.35.026.67.21.865.501L12 21l2.755-4.133a1.14 1.14 0 01.865-.501 48.172 48.172 0 003.423-.379c1.584-.233 2.707-1.626 2.707-3.228V6.741c0-1.602-1.123-2.995-2.707-3.228A48.394 48.394 0 0012 3c-2.392 0-4.744.175-7.043.513C3.373 3.746 2.25 5.14 2.25 6.741v6.018z" />
</svg>
</div>
<div class="discovery-title">
<span class="num">04</span>
<h3>思维链的伪装</h3>
</div>
</div>
<p class="en-title">The Disguise of Chain of Thought</p>
<p class="desc">
你的 AI 可能在撒谎。其输出的推理过程往往只是事后的"公关稿",用来掩盖它基于直觉和位置偏见的粗糙决策。
</p>
<p class="quote green-border">
Your AI might be lying. Its output reasoning is often just a post-hoc "press release" designed to cover up crude decisions based on intuition and positional bias.
</p>
</div>
</div>
</section>
<!-- Quote Section -->
<section class="quote-section">
<div class="quote-box animate-on-scroll">
<div class="quote-icon">
<svg xmlns="http://www.w3.org/2000/svg" fill="currentColor" viewBox="0 0 24 24">
<path d="M14.017 21v-7.391c0-5.704 3.731-9.57 8.983-10.609l.995 2.151c-2.432.917-3.995 3.638-3.995 5.849h4v10h-9.983zm-14.017 0v-7.391c0-5.704 3.748-9.57 9-10.609l.996 2.151c-2.433.917-3.996 3.638-3.996 5.849h3.983v10h-9.983z" />
</svg>
</div>
<p>
这不仅仅是技术瑕疵,更是对未来 <span class="highlight">AI 安全</span> 的严峻警示。
</p>
<p class="en">
This isn't just a technical glitch; it's a critical warning for the future of AI safety.
</p>
</div>
</section>
<!-- Stats Section -->
<section>
<div class="stats-grid animate-on-scroll">
<div class="stat-card pink-stat">
<div class="number">1,635</div>
<div class="label-cn">争议性问题</div>
<div class="label-en">Controversial Questions</div>
</div>
<div class="stat-card blue-stat">
<div class="number">15,058</div>
<div class="label-cn">证据文档</div>
<div class="label-en">Evidence Documents</div>
</div>
<div class="stat-card green-stat">
<div class="number">GroupQA</div>
<div class="label-cn">数据集名称</div>
<div class="label-en">Dataset</div>
</div>
</div>
</section>
<!-- Footer -->
<footer>
<div class="footer-line"></div>
<div class="footer-content">
<div class="footer-left">
<h3>卡内基梅隆大学</h3>
<p>Carnegie Mellon University</p>
</div>
<div class="footer-right">
<span>Research Paper</span>
<span class="arrow">›</span>
<span class="arxiv">arXiv:2601.06189</span>
</div>
</div>
<p class="footer-author">
Author: Atharv Naphade · LLM Research, Uncertainty, Reasoning, and AI Safety
</p>
</footer>
</div>
<script>
// Intersection Observer for scroll animations
const observerOptions = {
threshold: 0.1,
rootMargin: '0px 0px -50px 0px'
};
const observer = new IntersectionObserver((entries) => {
entries.forEach(entry => {
if (entry.isIntersecting) {
entry.target.style.opacity = '1';
entry.target.style.transform = 'translateY(0)';
}
});
}, observerOptions);
document.querySelectorAll('.animate-on-scroll').forEach(el => {
el.style.opacity = '0';
el.style.transform = 'translateY(30px)';
el.style.transition = 'opacity 0.8s ease, transform 0.8s ease';
observer.observe(el);
});
</script>
</body>
</html>
登录后可参与表态
讨论回复
0 条回复还没有人回复,快来发表你的看法吧!