Search papers, labs, and topics across Lattice.
Affiliation:
3
0
5
Hidden-state representation alignment in diffusion models isn't just an empirical heuristic鈥攊t is mathematically equivalent to score distillation operating directly within the latent representation space.
Recurrent models can extrapolate up to 128脳 beyond their training horizon simply by stabilizing backward-pass state credit without altering the forward computation.
Chimera achieves 7.3x compute efficiency over traditional models while enabling zero-shot extrapolation from short video clips to significantly longer sequences.
