Metraipolis › archive › dc_9db989b0de3d

CARVE: Verified Expansion for Variable-Length Generation in Diffusion Language Models — arXiv:2608.30922

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.30922
License (as archived)
CC0-1.0
Archived
2026-10-17T07:46:56.479Z
SHA-256 of the archived text
61afa9f8363267864f502ece7186884a93d451a0cf117ae028061774aeba2fab
Stable URL of this page
https://metraipolis.com/documento/dc_9db989b0de3d

How to cite

CARVE: Verified Expansion for Variable-Length Generation in Diffusion Language Models — arXiv:2608.30922 — arXiv (en), https://arxiv.org/abs/2608.30922 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_9db989b0de3d (Metraipolis archive dc_9db989b0de3d, sha256 61afa9f8363267864f502ece7186884a93d451a0cf117ae028061774aeba2fab)
BibTeX
@misc{dc9db989b0de3d,
  title = {CARVE: Verified Expansion for Variable-Length Generation in Diffusion Language Models — arXiv:2608.30922},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.30922},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_9db989b0de3d; sha256 61afa9f8363267864f502ece7186884a93d451a0cf117ae028061774aeba2fab},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.30922. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Masked diffusion language models predict tokens from a partially observed response canvas, enabling bidirectional conditioning and parallel token refinement. Yet standard masked-diffusion decoders use a rigid inference interface: the number of masked positions allocated to the answer is fixed before generation begins. Choosing this length is difficult. A short canvas can truncate reasoning or code, while a long canvas wastes computation and can perturb denoising. We introduce CARVE (Counterfactual-Aware Reveal with Verified Expansion), a training-free variable-length algorithm for masked diffusion LMs. Starting from a shorter canvas, CARVE can grow the response during decoding by inserting additional [MASK] positions. Rather than keeping every insertion, CARVE tests a candidate expanded canvas and asks a counterfactual question: would the model make similar predictions for the unresolved positions in the original canvas if the extra masked space were present? The inserted masks are kept only when they induce low Jensen-Shannon (JS) divergence on aligned unresolved positions. This makes length growth a verified stability decision rather than a pure confidence heuristic. CARVE applies without retraining to both full-canvas and blockwise diffusion decoders. Across code generation and mathematical reasoning benchmarks, CARVE consistently improves average performance over fixed-length baselines across all evaluated model families. Crucially, CARVE achieves these accuracy gains while reducing inference cost, reaching half the FLOPs of fixed-length decoding in some settings.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-01 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.30922 — 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 2026-10-17T07:28:26.633Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).