让大模型推理更省token，智能分配搜索资源。
Aligning Tree-Search Policies with Fixed Token Budgets in Test-Time Scaling of LLMs
- 根据剩余token动态调整搜索策略，先广搜后精修。
- 在数学和物理推理任务中均优于传统方法。
- 适合资源受限场景下提升大模型推理效率。
树搜索解码是大语言模型测试时扩展的有效方法，但实际部署常受限于每查询固定token预算，且不同场景差异显著。现有树搜索策略大多与预算无关，仅将预算视为终止条件，导致晚期过度分支或过早终止。我们提出预算引导的蒙特卡洛树搜索（BG-MCTS），其搜索策略随剩余预算动态调整：初期广泛探索，后期聚焦优化与答案完成，并减少浅层节点的分支。BG-MCTS在多个数学推理基准及一个物理推理基准上，对不同推理预算均显著优于预算无关的基线方法，适用于开源权重的大语言模型。
原文摘要 · Abstract (English)
Tree-search decoding is an effective form of test-time scaling for large language models (LLMs), but real-world deployment often imposes a fixed per-query token budget that varies across settings. Existing tree-search policies are largely budget-agnostic, treating the budget merely as a termination condition, thereby risking late-stage over-branching or premature termination. We propose Budget-Guided MCTS (BG-MCTS), a tree-search decoding algorithm that aligns its search policy with the remaining token budget: it starts with broad exploration, then prioritizes refinement and answer completion as the remaining budget decreases while reducing late-stage branching from shallow nodes. BG-MCTS consistently outperforms budget-agnostic tree-search baselines across inference budgets on mathematical reasoning benchmarks and an additional physics reasoning benchmark with open-weight LLMs.
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。