---
title: "重采样还是重路由？面向大语言模型的预算感知测试时模型选择"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2607.08665v1"
permalink: "https://feed.astralor.com/items/itm_86b7e16995ea79a5718d805b"
published: "2026-07-09T16:34:15.000Z"
score: 81
channel: "ai"
tags: ["arXiv","大型语言模型","成本优化","模型路由","测试时模型选择","重采样"]
---

# 重采样还是重路由？面向大语言模型的预算感知测试时模型选择

## AI 点评

> 这篇论文巧妙地抓住了LLM服务中一个被忽视的维度——同一查询预算下重采样与重路由的竞争关系。它提出的RoR策略简单而有效，对任何需要按查询付费的推理系统都有实际参考价值。对于AI基础设施团队和模型服务提供商来说，这是降低总成本同时保持高正确率的有力工具。值得关注的是，验证器质量是增益的关键门槛，这提醒我们在实际落地时需优先投资高质量验证器。

## 速览

本文提出一种预算感知的测试时模型选择方法，在每次查询的固定成本预算下，动态分配预算用于对当前已选定模型进行重采样（重新生成响应）或重路由（切换至其他模型），以最大化预期正确率。作者设计了基于边际正确率估计的在线分配策略（RoR），并在11个开源模型、4个基准上验证了其成本-质量Pareto前沿优于多种基线，特别在异质性最高的基准上增益最大。消融实验表明增益依赖于验证器质量，且在不同价格向量和标签无关验证器下结论稳健。

## 关键点

- 首次将重采样与重路由视为同一查询预算下的竞争性选择，提出预算感知的测试时模型选择问题。
- 提出在线Resample-or-Reroute (RoR) 分配策略，基于每单位成本的边际正确率估计进行预算分配。
- 在11个开源模型、4个难度不同的基准上实验，RoR优于单路由、单提交路由、Best-of-K、级联和随机分配等基线。
- 消融实验表明RoR的增益依赖于验证器质量，且在不同价格和标签无关验证器下结论仍然稳健。

## 全文

Routing among large language models (LLMs) trades response quality against serving cost, motivated by the reported gap between deployed routers and a per-instance oracle. Recent analysis shows that test-time resampling can recover per-instance selection headroom that no single-commit router captures; however, that guarantee holds only under an idealized oracle equipped with correctness labels and an unconstrained budget, neither of which a deployed system has. To the best of our knowledge, no previous work treats resampling the committed model and rerouting to an alternative model as competing uses of a single per-query cost budget. Therefore, this work formulates budget-aware test-time model selection: given a per-query budget and an imperfect verifier, allocate each unit of budget between resampling and rerouting so that expected correctness is maximized. An online resample-or-reroute (RoR) allocation policy driven by estimated marginal correctness per unit cost is proposed, and its behavior is grounded in the recoverability asymmetry between selection and sampling. Replay experiments on newly regenerated multi-draw correctness tensors from an eleven-model open-weight pool over four benchmarks of differing difficulty show that the proposed RoR policy attains a favorable cost-quality Pareto front relative to single-route, one-commit-router, budget-aware best-of-K, cascade, and random-allocation baselines for the tested pools, with the largest gains on the most heterogeneous benchmark; an ablation further shows the gains are verifier-gated, shrinking as verifier quality degrades, and robustness replays under a provider price vector and a label-free agreement verifier delineate where the conclusions carry over.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_86b7e16995ea79a5718d805b