arXiv cs.AI· Ziyi Chen, Yan Zhang, Jianhui Wei, Jianing Yin, Daoan Zhang·· 3 天前AI 评分34
大决策模型 LDMs：将大语言模型改造为单次前向决策器
Large Decision Models: Repurposing Large Language Models as Single-Pass Decision Makers
AI 导读
研究者提出大决策模型（LDMs），让语言模型通过一次 prefill 前向传播，直接输出运行时指定标签的显式概率。
正文 · 原文
Abstract:Large language models make decisions through parallel prefill and sequential decoding. Although effective, this approach incurs decoding costs and does not explicitly expose decision probabilities. Existing task-specific classifiers require per-task training and offer limited generalization to unseen decision schemas. We propose Large Decision Models (LDMs), repurposing language models to return explicit probabilities over labels specified at run time using one prefill forward pass. We first adapt models to this interface through Diverse Decision Supervision (DDS), covering varied domains, rules and labels. To further narrow the performance gap to a reasoning model, Self-Posed Reasoning Distillation (SPRD) adds tasks generated by a frozen teacher, filters them for answer consistency, and trains on label distributions obtained from repeated teacher reasoning. We evaluate on DecisionVista, pairing new-schema tests with held-out human-labelled sources. Across three base models, the recipe improves accuracy over a step-matched baseline by 1.8 and 1.5 percentage points, respectively, achieving a 182-fold speedup in latency over reasoning rollout on Gemma-4-31B. Reasoning remains stronger on the new-schema test, motivating a confidence router for uncertain decisions. On development sets, routing recovers a median 92% of reasoning's accuracy advantage while generating 27-58% of its tokens. Code is available at this https URL.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.15501 [cs.AI] |
| (or arXiv:2610.15501v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.15501 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jianhui Wei [view email]
[v1]
Sun, 11 Oct 2026 10:04:59 UTC (245 KB)
来源：arXiv cs.AI · arxiv.org