Metraipolis › archive › dc_dea0ae579f03

Rank Measures of Fine-Tuning Updates Need a Noise Baseline — arXiv:2610.22438

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.22438
License (as archived)
CC0-1.0
Archived
2026-10-19T08:48:17.520Z
SHA-256 of the archived text
6727b3f0af01f2e69c1c59c873645b3d80bd5df92c0f65849158a60dba8d7572
Stable URL of this page
https://metraipolis.com/documento/dc_dea0ae579f03

How to cite

Rank Measures of Fine-Tuning Updates Need a Noise Baseline — arXiv:2610.22438 — arXiv (en), https://arxiv.org/abs/2610.22438 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_dea0ae579f03 (Metraipolis archive dc_dea0ae579f03, sha256 6727b3f0af01f2e69c1c59c873645b3d80bd5df92c0f65849158a60dba8d7572)
BibTeX
@misc{dcdea0ae579f03,
  title = {Rank Measures of Fine-Tuning Updates Need a Noise Baseline — arXiv:2610.22438},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.22438},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_dea0ae579f03; sha256 6727b3f0af01f2e69c1c59c873645b3d80bd5df92c0f65849158a60dba8d7572},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.22438. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Studies of what fine-tuning does to a language model often summarise the weight update $\Delta W$ by its rank. We compare the rank measures used for this against the same measures on a random matrix of the same shape, and against two harder references: the pretrained weights, and matched fine-tunes without alignment content. Measures that count directions put our fine-tuning updates only a little below noise; stable rank, computed from the same singular values, puts them far below. Both are right: an update has a few strong directions on top of a long tail. The pattern holds for thirteen released pairs up to 8B parameters, and it is not a mark of alignment: with training held fixed, instruction data does not stand out from Wikipedia or code, DPO gives the same update whether its preference labels are real or half swapped at random (one model, modest preference strength), and scrambled text gives the most concentrated update of all. Applied to a recent claim that post-training is superficial because the effective rank of the weights barely changes, the baselines show that a random update of the same size leaves that rank almost unchanged while wrecking the model, and that an edit built on the effective rank of the update has no measurable effect in our tests. Rank measures of weight updates need a noise baseline, and claims about alignment need a control without alignment content.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.22438 — 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 2026-10-19T08:42:40.205Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).