Metraipolis › archive › dc_7b90602655ea
DepthGraft: Structural Regularization through Hierarchical Cross-Layer KV Reconstruction — arXiv:2611.03142
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.03142
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T11:11:29.096Z
- SHA-256 of the archived text
b0d8515a72ea588daf3f5d1f7848b14e1f4ea760b259d56be484ddf6a76bcdb7- Stable URL of this page
https://metraipolis.com/documento/dc_7b90602655ea
How to cite
DepthGraft: Structural Regularization through Hierarchical Cross-Layer KV Reconstruction — arXiv:2611.03142 — arXiv (en), https://arxiv.org/abs/2611.03142 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7b90602655ea (Metraipolis archive dc_7b90602655ea, sha256 b0d8515a72ea588daf3f5d1f7848b14e1f4ea760b259d56be484ddf6a76bcdb7)
BibTeX
@misc{dc7b90602655ea,
title = {DepthGraft: Structural Regularization through Hierarchical Cross-Layer KV Reconstruction — arXiv:2611.03142},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.03142},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7b90602655ea; sha256 b0d8515a72ea588daf3f5d1f7848b14e1f4ea760b259d56be484ddf6a76bcdb7},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.03142. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Scaling depth is a key route to stronger large language models, but Pre-LN architectures often suffer from depth-wise information dilution, limiting the effective use of deep layers. While cross-layer connectivity has proven effective at mitigating this problem, existing designs usually introduce additional computation or memory access. We revisit cross-layer KV sharing as an efficient form of such connectivity. Beyond cache compression, each KV reuse edge creates a direct gradient path from deep layers to source KV projections, turning KV sharing into an implicit structural regularizer for KV representation geometry. We provide a systematic analysis of this mechanism, showing that proper sharing topologies sharpen key retrieval directions while enriching the value content subspace, whereas existing adjacent or global-anchor designs do not fully exploit this effect. Motivated by this, we propose DepthGraft, a hierarchical KV-sharing architecture that reconstructs deep-layer KV from stride-aligned shallow-middle source pairs. DepthGraft distributes non-local gradient flow across source layers, while preserving prefill early completion and reducing KV cache by roughly one third. Experiments on dense and MoE models from 0.5B to 16B parameters show consistent downstream improvements under both GQA and MLA. Notably, on the 16B MoE model, DepthGraft achieves a 7.0% absolute improvement on MMLU-Pro, along with gains of 6.6% on CMMLU and 6.4% on C3. These results validate the potential of DepthGraft as a scalable design for large-scale pretraining. The project page is available at this https URL . 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/2611.03142 — 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 2026-11-03T10:59:52.360Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).