Metraipolis › archive › dc_309abec00a03

When Quantization Preserves Accuracy but Not Evidence: Explanation-Aware Post-Training Quantization for Medical LLMs — arXiv:2609.24799

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.24799
License (as archived): CC0-1.0
Archived: 2026-10-09T06:50:13.375Z
SHA-256 of the archived text: acb5348d8722d771eebd665e69a144e2bf689742cd67ffb872d2d6bbfddb4f5e
Stable URL of this page: https://metraipolis.com/documento/dc_309abec00a03

How to cite

When Quantization Preserves Accuracy but Not Evidence: Explanation-Aware Post-Training Quantization for Medical LLMs — arXiv:2609.24799 — arXiv (en), https://arxiv.org/abs/2609.24799 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_309abec00a03 (Metraipolis archive dc_309abec00a03, sha256 acb5348d8722d771eebd665e69a144e2bf689742cd67ffb872d2d6bbfddb4f5e)

BibTeX

@misc{dc309abec00a03,
  title = {When Quantization Preserves Accuracy but Not Evidence: Explanation-Aware Post-Training Quantization for Medical LLMs — arXiv:2609.24799},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.24799},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_309abec00a03; sha256 acb5348d8722d771eebd665e69a144e2bf689742cd67ffb872d2d6bbfddb4f5e},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.24799. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Post-training quantization (PTQ) enables efficient deployment of large language models, and PTQ methods are usually optimized and evaluated with generic reconstruction, perplexity, or answer accuracy. But in explanation-critical domains, preserving only the final answer may be insufficient, since users may also inspect generated rationales to judge whether a prediction is trustworthy. We study this issue in medical multiple-choice question answering, where rationales should provide evidence that supports the selected answer. We propose an explanation-aware objective for transformation-based PTQ. Our method builds an offline faithfulness cache from full-precision teacher rationales and uses it during optimization to preserve answer-supporting evidence tokens and evidence-conditioned answer behavior. We instantiate it on OSTQuant under W4A4KV4 quantization and evaluate four 7B--8B medical and instruction-tuned LLMs on MedExQA, MedExpQA, and ChallengeClinicalQA. While a same-calibration OSTQuant baseline preserves task accuracy, it can substantially weaken answer-supporting rationales. Our objective is to preserve the full-precision model's answer-supporting behavior rather than improve gold-label accuracy, and our method better preserves the full-precision model's answer behavior and rationale-to-answer support. These results suggest that PTQ for explanation-critical settings should evaluate preservation of answer-supporting evidence, not only answer accuracy. Code and evaluation scripts are available at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-22 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.24799 — 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 2026-10-09T06:48:45.062Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-22?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).