Search papers, labs, and topics across Lattice.
Affiliation:
2
0
3
1
Reward hacking is mitigated and sample quality preserved in few-step discrete diffusion models by replacing intractable KL penalties with a novel inverse-distillation regularization.
Non-factorized diffusion language models achieve superior few-step generation by leveraging a discrete latent space learned via Mixture-of-Experts routing, bypassing posterior collapse issues.
