Quantifying Pretraining Variance: Floating-Point Order Effects Nearly as Large as Data Seeds

AndrewLampinen · x · 2026-08-19

A new blog post quantifies sources of variance in LLM pretraining. Beyond data and initialization seeds, floating-point arithmetic order (e.g., from sharding differences) also affects results, with effects nearly as large as data seeds.

Original post →

More from Research

Research channel →