Metraipolis › archive › dc_db4e13564fde

X-MoD: Practical Scaling Laws for Sparse-Depth Routing Beyond Mixture-of-Depths — arXiv:2609.34212

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.34212
License (as archived): CC0-1.0
Archived: 2026-10-07T21:54:10.124Z
SHA-256 of the archived text: 7b7126541dd430b1cc9a2a1c85b98ad15f271e05c910b3cb88b1120f0db8f0b4
Stable URL of this page: https://metraipolis.com/documento/dc_db4e13564fde

How to cite

X-MoD: Practical Scaling Laws for Sparse-Depth Routing Beyond Mixture-of-Depths — arXiv:2609.34212 — arXiv (en), https://arxiv.org/abs/2609.34212 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_db4e13564fde (Metraipolis archive dc_db4e13564fde, sha256 7b7126541dd430b1cc9a2a1c85b98ad15f271e05c910b3cb88b1120f0db8f0b4)

BibTeX

@misc{dcdb4e13564fde,
  title = {X-MoD: Practical Scaling Laws for Sparse-Depth Routing Beyond Mixture-of-Depths — arXiv:2609.34212},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.34212},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_db4e13564fde; sha256 7b7126541dd430b1cc9a2a1c85b98ad15f271e05c910b3cb88b1120f0db8f0b4},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.34212. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Mixture-of-Depths (MoD) enables conditional computation across Transformer depth by routing only a subset of tokens through selected layers, but its original one-sparse--one-dense alternation tightly couples total capacity to active capacity and limits sparse-depth scaling. We introduce X-MoD, a scalable sparse-depth architecture that decouples token sparsity from anchor stride, allowing total parameter count to grow while keeping active-equivalent capacity nearly fixed. To make deep sparse routing trainable, X-MoD combines dense anchors with variance-scaled layer-wise gating and depth-wise token balancing. To make this regime analyzable and usable, we formulate sparse-depth routing as a conditional architecture-design problem: given compute, context length, and active-equivalent backbone size, how should the routing configuration be chosen? We develop a practical scaling-law framework by fitting X-MoD relative to FLOP-matched dense baselines, yielding an interpretable law that decomposes performance into sparse-capacity gain, sparse-context correction, and anchor-stride interaction. The law predicts validation loss across routing configurations and reveals how context length, model scale, and anchor stride shape sparse-depth performance. We validate the architecture and law through pretraining sweeps, held-out scaling-law prediction, ablations, downstream evaluations, and comparisons with Dense, MoD, and representative MoE baselines.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.CL).
Categorie arXiv: Machine Learning (cs.LG) ; Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.34212 — 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 2026-10-07T21:27:08.800Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).