pith. sign in

Fast-decoding diffusion language models via progress-aware confidence schedules

2 Pith papers cite this work. Polarity classification is still indexing.

2 Pith papers citing it

citation-role summary

background 1

citation-polarity summary

fields

cs.CL 1 cs.LG 1

years

2026 2

verdicts

UNVERDICTED 2

roles

background 1

polarities

background 1

representative citing papers

Multi-Token Residual Prediction

cs.LG · 2026-05-12 · unverdicted · novelty 7.0

MRP predicts logit residuals from hidden states to support dependency-aware multi-token denoising in a single forward pass for diffusion language models, yielding up to 1.42× lossless speedup on SDAR models.

citing papers explorer

Showing 2 of 2 citing papers.

  • Multi-Token Residual Prediction cs.LG · 2026-05-12 · unverdicted · none · ref 25

    MRP predicts logit residuals from hidden states to support dependency-aware multi-token denoising in a single forward pass for diffusion language models, yielding up to 1.42× lossless speedup on SDAR models.

  • TAD: Temporal-Aware Trajectory Self-Distillation for Fast and Accurate Diffusion LLM cs.CL · 2026-05-10 · unverdicted · none · ref 13

    TAD improves the accuracy-parallelism trade-off in diffusion LLMs via temporal-aware self-distillation that applies hard labels to soon-to-be-decoded tokens and soft supervision to future tokens.