Pengxiao Lin

Uncritical tokens are 'critical' in pretraining: the implicit regularization effect of Next token prediction

Identity Bridge: Enabling Implicit Reasoning via Shared Latent Memory

Achilles' Heel of Mamba: Essential difficulties of the Mamba architecture demonstrated by synthetic data

M$^2$M: LEARNING CONTROLLABLE MULTI OF EXPERTS AND MULTI-SCALE OPERATORS ARE THE PARTIAL DIFFERENTIAL EQUATIONS NEED