Report
OLMo3-32B reportedly dropped to 0% on an arithmetic diagnostic, then rebounded
The poster calls the swing “mode-hopping” and says validation loss decreased smoothly throughout.
TLDR
A post says OLMo3-32B’s arithmetic diagnostic accuracy reached 81% at 2.17 trillion pre-training tokens, fell to 0% at 2.19 trillion, then rose to 81.7% at 2.21 trillion. The poster calls it “mode-hopping” and says validation loss decreased smoothly throughout.
Combined views
—
2 Sources, first seen ago
— likes— comments— saves— reposts
