Large Language Bayes 不是重参数化不变的

1. 12:00
   arXiv · Machine Learning Theory

   Large Language Bayes 不具备重参数化不变性

   论文证明大语言贝叶斯（LLB）基于指数化证据界的加权依赖于模型写法，而对数边际似然具有重参数化不变性、证据界不然。在八所学校中心与非中心程序差异仅 5.7×10⁻¹⁴，但权重相差 6.1×；重要性加权降至 2.2×，复现 LLB 全协方差高斯匹配后验矩仍剩 1.9×（八学校）和 8.9×（64 维）。

Current heat 9·Comparable peak 10（Oct 2）·Comparable change over 24 hours –

The trend compares only the same participants observed continuously; its range may be smaller than the current heat count. Move or click on the chart to inspect hourly heat; use the left and right arrow keys to switch.