Metraipolis › archive › dc_fb4504e9323f

When PTQ Calibration Helps or Hurts: Dose-Driven Collapse and Validation-Gated Abstention in Quantized Text Embedders — arXiv:2610.27397

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.27397
License (as archived)
CC0-1.0
Archived
2026-10-21T07:06:31.293Z
SHA-256 of the archived text
4a7d37585e8ec931c2556362e953f0a84296f3244ed0c09dadf2a4641ff37ea8
Stable URL of this page
https://metraipolis.com/documento/dc_fb4504e9323f

How to cite

When PTQ Calibration Helps or Hurts: Dose-Driven Collapse and Validation-Gated Abstention in Quantized Text Embedders — arXiv:2610.27397 — arXiv (en), https://arxiv.org/abs/2610.27397 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fb4504e9323f (Metraipolis archive dc_fb4504e9323f, sha256 4a7d37585e8ec931c2556362e953f0a84296f3244ed0c09dadf2a4641ff37ea8)
BibTeX
@misc{dcfb4504e9323f,
  title = {When PTQ Calibration Helps or Hurts: Dose-Driven Collapse and Validation-Gated Abstention in Quantized Text Embedders — arXiv:2610.27397},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.27397},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fb4504e9323f; sha256 4a7d37585e8ec931c2556362e953f0a84296f3244ed0c09dadf2a4641ff37ea8},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.27397. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Learned post-training quantization (PTQ) calibration fits quantizer parameters on a small calibration set, and for a text embedder the natural supervision is retrieval relevance. We ask two questions that are usually conflated: how far to calibrate, and whether to calibrate at all. Across three embedders (e5-base-v2, BGE-M3, EmbeddingGemma-300m) at 4-bit weight quantization, 161 calibration runs, the outcome of a fixed procedure ranges from a gain to catastrophic degradation, and the driver is the dose of the update, measured as the functional displacement D_func of the quantized model from its uncalibrated start. On EmbeddingGemma the same learning-rate grid gives +1.47 to -71.53 points of in-domain NDCG@10 change; a label-free cosine objective collapses at the same dose (-69.52 vs -71.63), so the collapse is a property of learnable calibration rather than of retrieval supervision. Objective sensitivity is heterogeneous: on e5, ranking calibration is substantially more robust than cosine at matched displacement (+14.16 +/- 4.34 pp, 95% t-CI [+3.37, +24.96], three paired slates), whereas on BGE-M3 the ordering is directionally opposite but unresolved ([-5.18, +2.38]). Dose is the main axis but not a sufficient statistic: equal D_func reached along different paths differs by up to 1.61 pp. We therefore treat calibration as an optional intervention. A pre-registered rule that accepts a checkpoint only when a per-query paired confidence interval on a 25% calibration-validation split clears the uncalibrated baseline was run on 6 independent slates and five arms. It never selected a checkpoint below its slate's baseline (0 harmful selections in 30 decisions), retained 8 of 11 beneficial updates, and paid a mean oracle regret of 1.52 pp with one abstention that forfeited 11.13 pp, a safety-opportunity trade-off, not a free improvement. We report what remains open, including label efficiency.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.27397 — 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 2026-10-21T07:05:43.404Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).