Metraipolis › archive › dc_03cd5ebcf074

Counterfactual Pairwise Augmentation Improves Small Pointwise LLM Judges — arXiv:2610.41149

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.41149
License (as archived): CC0-1.0
Archived: 2026-11-03T07:31:22.596Z
SHA-256 of the archived text: 991efcfd019d20c0af0c9cb57fa3233383b5cb9400ac7699d70bbdea3c01501e
Stable URL of this page: https://metraipolis.com/documento/dc_03cd5ebcf074

How to cite

Counterfactual Pairwise Augmentation Improves Small Pointwise LLM Judges — arXiv:2610.41149 — arXiv (en), https://arxiv.org/abs/2610.41149 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_03cd5ebcf074 (Metraipolis archive dc_03cd5ebcf074, sha256 991efcfd019d20c0af0c9cb57fa3233383b5cb9400ac7699d70bbdea3c01501e)

BibTeX

@misc{dc03cd5ebcf074,
  title = {Counterfactual Pairwise Augmentation Improves Small Pointwise LLM Judges — arXiv:2610.41149},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.41149},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_03cd5ebcf074; sha256 991efcfd019d20c0af0c9cb57fa3233383b5cb9400ac7699d70bbdea3c01501e},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41149. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Agentic systems increasingly gate individual steps with a pointwise judge that scores a response against a criterion. Small language models are attractive for this role because they can be served and specialized efficiently. Such a judge must locate the criterion's decision boundary, but pointwise labels under-determine that boundary in fine-tuning: labeled responses to different queries differ along many dimensions at once, so they do not reveal which features decide the label. We introduce CPA (Counterfactual Pairwise Augmentation), which makes the deciding difference explicit in training. For each pointwise item, CPA synthesizes and audits a counterfactual response to the same query intended to carry the opposite label. It frames the accepted pair as a pairwise task asking which of the two responses satisfies the criterion. The judge is fine-tuned jointly on both pointwise and pairwise datasets and deployed in pointwise mode only. A second-order analysis shows when this should outperform the prevailing strategy of merging separately trained pointwise and pairwise experts: merging moves every parameter direction toward the pairwise expert by the same fraction, whereas joint training can hold the pointwise optimum wherever the pairwise loss is flat and absorb its curvature where it is sharp. This is the regime that CPA's same-source, counterfactually minimal pairwise view is designed to produce. On separate Qwen3-8B LoRA adapters for six binary criteria, CPA improves criterion-averaged pointwise macro-F1 by 2.5 points over pointwise-only training and outperforms model merging on all six criteria at every tested merge weight. It also surpasses the strongest zero-shot frontier judge we evaluate, Gemini 3.6 Flash, by 11.8 points.

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/2610.41149 — 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 2026-11-03T07:29:33.602Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).