Metraipolis › archive › dc_a67c90d2d457

DBLAST: Dependent Block Drafting for Stochastic Speculative Decoding — arXiv:2608.05448

Source: arXiv (en)
Original: https://arxiv.org/abs/2608.05448
License (as archived): CC0-1.0
Archived: 2026-10-25T21:27:08.782Z
SHA-256 of the archived text: c717ed2f2e10c4c8528a84fd8f68a3f7bf74dd4ad1e12a8c7aedbe16eafdb1b5
Stable URL of this page: https://metraipolis.com/documento/dc_a67c90d2d457

How to cite

DBLAST: Dependent Block Drafting for Stochastic Speculative Decoding — arXiv:2608.05448 — arXiv (en), https://arxiv.org/abs/2608.05448 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a67c90d2d457 (Metraipolis archive dc_a67c90d2d457, sha256 c717ed2f2e10c4c8528a84fd8f68a3f7bf74dd4ad1e12a8c7aedbe16eafdb1b5)

BibTeX

@misc{dca67c90d2d457,
  title = {DBLAST: Dependent Block Drafting for Stochastic Speculative Decoding — arXiv:2608.05448},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.05448},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a67c90d2d457; sha256 c717ed2f2e10c4c8528a84fd8f68a3f7bf74dd4ad1e12a8c7aedbe16eafdb1b5},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.05448. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Speculative decoding accelerates large language models' inference by using a lightweight drafter to propose multiple future tokens and a target model to verify them. While recent block and diffusion-style drafters can predict several positions in a single pass, their training and sampling procedures are typically optimized for greedy decoding or assume that positions in the draft block are conditionally independent. This assumption becomes brittle in non-greedy speculative decoding, where the target distribution is deliberately stochastic and multiple continuations become plausible. We study this mismatch for block diffusion drafters and show that the accepted draft length degrades as the entropy of the target sampling distribution increases. We propose a dependent block drafter based on a low-rank latent mixture over token positions, complemented by an acceptance-oriented training objective that directly targets the expected verified length. Experiments with Qwen3-4B and Qwen3-8B on GSM8K, MT-Bench, HumanEval, and creative-writing benchmarks show that our approach, namely DBLast, consistently improves accepted length over independent block sampling, especially in higher-entropy decoding regimes.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-07 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2608.05448 — 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 2026-10-25T21:24:38.191Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-07?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).