Bayesian evidence recovers dimensionality where held-out loss fails — at equal n

BlackHC · x · 2026-09-09

Continuing the thread: the growing signal is how Bayesian evidence recovers dimensionality in Bayesian PCA where held-out loss struggles (Lotfi et al. 2022).

Key caveat: the identity holds at equal n. Simpler models spend fewer FLOPs per example, and scaling laws put compute on the x-axis, where areas mean something else.

Related event: Bayesian Model Selection Explains Why Proxy-Scale Winners Fail at Target Scale(10 posts)→

Original post →

More from Research

Research channel →