通过熵调控实现高效模型路由，兼顾精度与计算成本。
EntroRouter: Learning Efficient Model Routing via Entropy Regulation
- 采用熵调控机制，在单轮中解耦推理与路由过程。
- 保留最强专家98.3%精度，计算成本降低48.25%。
- 适合需要高效推理的部署场景，如边缘计算。
模型路由通过在不同能力的模型间选择，平衡解决方案的准确率与计算开销。尽管近期多轮框架实现了推理与规划的交替，我们发现一种称为‘信任区域坍塌’的结构缺陷。研究显示，推理与路由的深度耦合，加上稀疏监督下强预训练先验的主导作用，导致可信赖的局部最优解中，高能力专家被系统性压制。为此，我们提出$ extbf{EntroRouter}$，一个单轮路由框架，将熵调控作为核心目标。首先通过软监督初始化策略，拟合合适模型的分布，建立高熵先验以促进探索；随后利用离线能力估计构建软锚点，稳定强化学习，实现安全信任区域内可控的熵收缩。大量实验表明，EntroRouter在保持最强专家98.3%准确率的同时，将计算成本降低了48.25%。
原文摘要 · Abstract (English)
Model routing balances solution accuracy and computational cost by selecting among models of varying capabilities. While recent multi-round frameworks interleave reasoning and planning, we identify a structural failure mode termed Trust Region Collapse. We demonstrate that the deep coupling of reasoning and routing, exacerbated by the dominance of strong pre-training priors under sparse supervision, leads to degenerate local optima where capable experts are systematically suppressed. To decouple these processes, we propose $\textbf{EntroRouter}$, a single-round routing framework that treats entropy regulation as a core objective. We first initialize the policy via Soft Supervision, fitting a distribution of suitable models to establish a high-entropy prior for exploration. Subsequently, we stabilize Reinforcement Learning using a Soft Anchor, which utilizes offline capability estimates to orchestrate controlled entropy contraction within a safe trust region. Extensive experiments demonstrate that EntroRouter retains 98.3% of the strongest expert's accuracy while reducing computational costs by 48.25%.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。