根据答案分歧动态选策略，提升数学推理模型准确性并省计算
When to Vote, When to Rewrite: Disagreement-Guided Strategy Routing for Test-Time Scaling
- 用输出分歧判断题目难易，决定采用投票、重写还是直接放行
- 在7个数学数据集上准确率提升3%~7%，且采样成本更低
- 无需训练，适配多种模型，适合追求效率与精度的部署场景
大型推理模型在数学推理任务上表现强劲，但在难题上仍不可靠。现有测试时扩展方法如重复采样、自修正和树搜索虽能提升性能，但计算开销大，且在难题上收益递减。我们发现输出分歧与题目难度及预测正确性高度相关，可作为测试时实例级策略选择的有效信号。基于此，提出一种无需训练的框架，将测试时扩展转化为实例级路由问题：根据分歧程度动态选择不同策略——一致结果直接输出，中等分歧采用多数投票，高度模糊则进行重写重构。在七个数学基准和三种模型上的实验表明，该方法在准确率提升3%~7%的同时，显著降低采样成本。
原文摘要 · Abstract (English)
Large Reasoning Models (LRMs) achieve strong performance on mathematical reasoning tasks but remain unreliable on challenging instances. Existing test-time scaling methods, such as repeated sampling, self-correction, and tree search, improve performance at the cost of increased computation, yet often exhibit diminishing returns on hard problems. We observe that output disagreement is strongly correlated with instance difficulty and prediction correctness, providing a useful signal for guiding instance-level strategy selection at test time. Based on this insight, we propose a training-free framework that formulates test-time scaling as an instance-level routing problem, rather than allocating more computation within a single strategy, dynamically selecting among different scaling strategies based on output disagreement. The framework applies lightweight resolution for consistent cases, majority voting for moderate disagreement, and rewriting-based reformulation for highly ambiguous instances. Experiments on seven mathematical benchmarks and three models show that our method improves accuracy by 3% - 7% while reducing sampling cost compared to existing approaches.
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。