Researchers · Interviewed at COLM 2026
Carnegie Mellon University

COLM 2026 · October 5, 2026 · 32 min
Do all language models learn things in the same order - even when they are trained on different data? Emmy Liu (CMU) tracked 91 skills across four model families through the first trillion tokens of pre-training, and the answer is: largely, yes.
Work is presented at COLM 2026, the Third Annual Conference on Language Modeling.