Metraipolis › archive › dc_bc3915882de7

TeCoP: Accelerating Diffusion Large Language Model Decoding via Tentative Context Pipelining — arXiv:2610.27583

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.27583
License (as archived)
CC0-1.0
Archived
2026-10-21T07:08:57.916Z
SHA-256 of the archived text
d682d691bdd296d7b5e5fdb9b5270e5b0627f11a2d7a61416323ed3881d8320e
Stable URL of this page
https://metraipolis.com/documento/dc_bc3915882de7

How to cite

TeCoP: Accelerating Diffusion Large Language Model Decoding via Tentative Context Pipelining — arXiv:2610.27583 — arXiv (en), https://arxiv.org/abs/2610.27583 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bc3915882de7 (Metraipolis archive dc_bc3915882de7, sha256 d682d691bdd296d7b5e5fdb9b5270e5b0627f11a2d7a61416323ed3881d8320e)
BibTeX
@misc{dcbc3915882de7,
  title = {TeCoP: Accelerating Diffusion Large Language Model Decoding via Tentative Context Pipelining — arXiv:2610.27583},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.27583},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bc3915882de7; sha256 d682d691bdd296d7b5e5fdb9b5270e5b0627f11a2d7a61416323ed3881d8320e},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.27583. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Diffusion large language models (dLLMs) predict multiple token positions per forward pass, yet block-wise decoders serialize generation and repeatedly recompute stable predictions. We attribute this inefficiency to coupling a prediction's readiness for irreversible commitment with its utility as context. We introduce Tentative Context Pipelining (TeCoP), a training-free decoder that decouples these roles through staggered block-parallel pipelining and draft-and-verify. TeCoP reuses stable but uncommitted predictions as provisional cross-block context, with commitment governed by confidence and stability checks. Across four LLaDA and Dream dLLMs on GSM8K, Minerva MATH, and MBPP, TeCoP reduces number of forward evaluations by up to 87% and achieves up to 7.50-fold throughput without KV caching and 11.77-fold with caching over the uncached original decoder. It outperforms Fast-dLLM and DSB in throughput across all 12 model-benchmark settings, with and without caching, while keeping average cached accuracy within one percentage point of the original decoder.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.27583 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-21T07:05:58.121Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).