Newest papers

Commit or Abstain? System One Decision Models, LLMs, and Missing Legal Authority

M. Mikail Demir · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.29961 · cited by 0 works (OpenAlex) · open-access PDF

Spanish

Dado el enorme número de reclamaciones que se deben comprobar, la IA legal verificable requiere modelos escalables y rentables que respondan a las necesidades prácticas de la ley. Los modelos de decisión del Sistema Uno se construyen en modelos de lenguaje grande pero, en lugar de generar texto, devuelven una probabilidad calibrada para cada opción. Esto es adecuado para la verificación de la autoridad legal, donde se cierra la cuestión: una posesión dada, o ninguno de ellos. Testamos si tales modelos se abstengan cuando la autoridad correcta está ausente. Extendemos CaseHOLD con una condición de abstención en la que la posesión correcta se sustituye en 1.670 de 5.314 artículos, y comparamos cuatro modelos de decisión con seis LLM generativos. Cuando ninguna opción es correcta, los LLM generativos todavía eligen uno en 38-86%

Machine translation (Meta M2M100); the English original follows.

Abstract

Given the enormous number of claims to be checked, verifiable legal AI requires scalable, cost-effective models that answer the practical needs of law. System One decision models are built on large language models but, instead of generating text, return a calibrated probability for each option. This suits the verification of legal authority, where the question is closed: a given holding, or none of them. We test whether such models abstain when the correct authority is missing. We extend CaseHOLD with an abstention condition in which the correct holding is replaced in 1,670 of 5,314 items, and compare four decision models with six generative LLMs. When no option is correct, generative LLMs still pick one in 38-86% of items; the most cautious decision model does so in 30%. On standard CaseHOLD, a decision model matches the best generative LLM (88.1% vs. 88.2%) at 1/55 of the cost. Accepting that LLM's answer only when a decision model agrees halves its wrong citations, and distilling a decision model's probabilities into Qwen3-8B, without labels, yields a well-calibrated local model. Legal AI should be judged by its wrong citations, not by accuracy alone.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Commit, Computer science, Generative grammar, Decision model, Verifiable secret sharing, Decision analysis

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange