<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Empirical on 中文AI术语词典</title><link>https://terms-en.ai-term-hub.com/zh/tags/empirical/</link><description>Recent content in Empirical on 中文AI术语词典</description><generator>Hugo</generator><language>zh-cn</language><lastBuildDate>Sat, 18 Jul 2026 11:44:45 +0000</lastBuildDate><atom:link href="https://terms-en.ai-term-hub.com/zh/tags/empirical/index.xml" rel="self" type="application/rss+xml"/><item><title>神经缩放定律</title><link>https://terms-en.ai-term-hub.com/zh/terms/neural_scaling_law/</link><pubDate>Sat, 18 Jul 2026 11:28:12 +0000</pubDate><guid>https://terms-en.ai-term-hub.com/zh/terms/neural_scaling_law/</guid><description>&lt;h2 id="definition">Definition&lt;/h2>
&lt;p>神经缩放定律描述了模型性能与其规模（包括数据集大小、参数量和计算预算）之间的可预测的幂律关系。这些定律表明，随着规模的扩大，模型性能会按特定规律提升。&lt;/p>
&lt;h3 id="summary">Summary&lt;/h3>
&lt;p>基于数据、参数或计算量增加来预测模型性能提升的经验关系。&lt;/p>
&lt;h2 id="key-concepts">Key Concepts&lt;/h2>
&lt;ul>
&lt;li>幂律关系&lt;/li>
&lt;li>参数缩放&lt;/li>
&lt;li>数据效率&lt;/li>
&lt;li>计算最优训练&lt;/li>
&lt;/ul>
&lt;h2 id="use-cases">Use Cases&lt;/h2>
&lt;ul>
&lt;li>大语言模型的资源规划&lt;/li>
&lt;li>模型架构选择&lt;/li>
&lt;li>性能预测&lt;/li>
&lt;/ul>
&lt;h2 id="related-terms">Related Terms&lt;/h2>
&lt;ul>
&lt;li>&lt;a href="https://terms-en.ai-term-hub.com/en/terms/chinchilla-optimization-chinchilla%E4%BC%98%E5%8C%96/">Chinchilla optimization (Chinchilla优化)&lt;/a>&lt;/li>
&lt;li>&lt;a href="https://terms-en.ai-term-hub.com/en/terms/loss-scaling-%E6%8D%9F%E5%A4%B1%E7%BC%A9%E6%94%BE/">Loss scaling (损失缩放)&lt;/a>&lt;/li>
&lt;li>&lt;a href="https://terms-en.ai-term-hub.com/en/terms/emergent-abilities-%E6%B6%8C%E7%8E%B0%E8%83%BD%E5%8A%9B/">Emergent abilities (涌现能力)&lt;/a>&lt;/li>
&lt;li>&lt;a href="https://terms-en.ai-term-hub.com/en/terms/compute-budget-%E8%AE%A1%E7%AE%97%E9%A2%84%E7%AE%97/">Compute budget (计算预算)&lt;/a>&lt;/li>
&lt;/ul></description></item></channel></rss>