LayWM uses 1,024 dimensions for its latent space representations, requiring 1,024 random 1D projections to verify through the Cramér-Wold theorem that the representation maintains Gaussian distribution.

factualpending

Speaker

Unidentified Speaker 1 — LeWorldModel: Stable End-to-End JEPA World Models From Pixe… [8OYwEWwpk4M]

Evidence Quote

1,024 dimensions in this case... It projects the data onto 1,024 random 1D lines.

Source

LeWorldModel: Stable End-to-End JEPA World Models From PixelsXiaol.x
Created: 8/12/2026, 10:24:38 PM

My Notes

Loading notes...