Metraipolis › archive › dc_299244ad07ad
Spectral Rank Certification for Foundation Model Adapters — arXiv:2608.15351
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.15351
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-18T19:40:13.916Z
- SHA-256 of the archived text
2ea9c37fe97d3f0f934daa4a6ed0c75c31c3bc6741e9c0018db6396b97861bbd- Stable URL of this page
https://metraipolis.com/documento/dc_299244ad07ad
How to cite
Spectral Rank Certification for Foundation Model Adapters — arXiv:2608.15351 — arXiv (en), https://arxiv.org/abs/2608.15351 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_299244ad07ad (Metraipolis archive dc_299244ad07ad, sha256 2ea9c37fe97d3f0f934daa4a6ed0c75c31c3bc6741e9c0018db6396b97861bbd)
BibTeX
@misc{dc299244ad07ad,
title = {Spectral Rank Certification for Foundation Model Adapters — arXiv:2608.15351},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.15351},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_299244ad07ad; sha256 2ea9c37fe97d3f0f934daa4a6ed0c75c31c3bc6741e9c0018db6396b97861bbd},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.15351. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Nominal LoRA rank is a design parameter; calibrated spectral evidence is a separate inferential quantity. This article develops a finite-sample framework for inferring effective rank structure in public foundation-model adapters. The theoretical core is an exact chi-square divergence for the fixed-dimensional Gaussian rank-one reference experiment, with an unknown signal direction integrated under a rotation-invariant reference prior. The resulting series yields a computable finite-sample Le Cam bound at concrete layer sizes, an explicit remainder bound for numerical truncation, and the rectangular Baik-Ben Arous-Peche (BBP) limit. A compact-manifold Laplace expansion shows that finite-sample likelihood evidence also depends on leading spectral gaps through the factor $s_1^{|m-n|}\prod_{i\ge2}(s_1^2-s_i^2)$, motivating joint calibration of clustered singular values. Building on these results, we introduce an empirical-null workflow for PEFT LoRA adapters: factor reconstruction, Monte Carlo $p$-values, stagewise and block testing, and module-wise and corpus-level BH reporting. In an audit of 26 public adapters, 684 modules, six architecture families, and 31,770 public-checkpoint spectra rows, calibrated effective rank is typically much smaller than nominal rank and differs systematically from 95\% energy retention. A measured RoBERTa-RTE slice on $n=24$ examples illustrates the measurement path from calibrated ranks to task evaluation, without treating the slice as a utility study. The main empirical finding is that calibrated effective rank is usually far below nominal rank, and that energy retention and statistical surprise answer different questions.
Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-18 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG) ; Methodology (stat.ME).
Pagina del lavoro: https://arxiv.org/abs/2608.15351 — 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 2026-10-18T19:02:44.085Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-08-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).