Models

GLM-5.2

A large language model used to evaluate the end-to-end performance of DoSA.

1 story, first covered on Oct 15, 2026.

Every story about GLM-5.2 (1)

$A figure from the paper \\\"DoSA: Dots Sparse Attention\\\".$ How it worksNew attention method speeds up AI models with mathematical shortcutsHerald of AIFigure: Meng et al., arXiv:2610.13990, CC BY 4.0

New attention method speeds up AI models with mathematical shortcuts

A paper introduces DoSA, which finds important tokens faster using bounds instead of full calculations.

Oct 15, 2026