PolyPythias: Stability and Outliers across Fifty Language Model Pre-Training Runs

Open Webpage

Oskar van der Wal, Pietro Lesci, Max Müller-Eberstein, Naomi Saphra, Hailey Schoelkopf, Willem H. Zuidema, Stella Biderman

Published: 01 Jan 2025, Last Modified: 20 May 2025ICLR 2025EveryoneRevisionsBibTeXCC BY-SA 4.0