Metraipolis › archive › dc_720bb284d1fb

Honest Small Decision Models for Finance: Scale, Specialization Collapse, and Mixed-Corpus Remediation — arXiv:2611.03755

Source: arXiv (en)
Original: https://arxiv.org/abs/2611.03755
License (as archived): CC0-1.0
Archived: 2026-11-04T12:39:18.967Z
SHA-256 of the archived text: 06a311bcf9e705f64902f5f5c606d1532a66f8c86629ba76291c95beaac86b77
Stable URL of this page: https://metraipolis.com/documento/dc_720bb284d1fb

How to cite

Honest Small Decision Models for Finance: Scale, Specialization Collapse, and Mixed-Corpus Remediation — arXiv:2611.03755 — arXiv (en), https://arxiv.org/abs/2611.03755 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_720bb284d1fb (Metraipolis archive dc_720bb284d1fb, sha256 06a311bcf9e705f64902f5f5c606d1532a66f8c86629ba76291c95beaac86b77)

BibTeX

@misc{dc720bb284d1fb,
  title = {Honest Small Decision Models for Finance: Scale, Specialization Collapse, and Mixed-Corpus Remediation — arXiv:2611.03755},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.03755},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_720bb284d1fb; sha256 06a311bcf9e705f64902f5f5c606d1532a66f8c86629ba76291c95beaac86b77},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.03755. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Financial workflows need fast, calibrated, auditable decisions, not just generated text. We train System-1 decision models, small language models fine-tuned to answer typed questions (choice/score/noul) with a probability in a single forward pass, across two model families (K2-Horizon and Qwen3.5/3.8) and nine rungs from 0.8B to 32B parameters, using one frozen LoRA recipe on a public decision benchmark and two new finance corpora: a programmatic-gold policy corpus (rule-reading with exact labels) and a reality-labeled outcome corpus (19,640 realized forward-return questions in walk-forward splits, plus a sealed 2026 test of 1,360 questions). Narrow fine-tuning costs general decision competence 9-45 percentage points, a data-mixture effect that shrinks with model scale and is associated with narrow-gradient volume. On reality-labeled questions, every trained rung from 0.8B to 32B lands at the horizon-conditioned prior baseline (accuracy 0.37-0.38, no detectable conditional edge under our controls); the models respond with flat, well-calibrated probabilities that track per-horizon base rates: naturally honest uncertainty, with a mid-scale calibration sweet spot (ECE 0.026 at 2B, degrading to 0.043-0.044 at 9-32B under prior drift). A mixed-corpus remediation carries all competencies in one adapter at modest cost: it matches or slightly exceeds the generalist specialist on the public benchmark while holding the policy ceiling and near-specialist outcome calibration, yielding a deployable artifact with per-domain temperature regimes and a band-dispatch escalation analysis: sharp, dispatchable bands where rules determine answers; flat, review-routing bands where reality does not.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG) ; Statistical Finance (q-fin.ST).
Pagina del lavoro: https://arxiv.org/abs/2611.03755 — 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 2026-11-04T12:29:47.242Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).