StepBack：面向高效测试时推理的步骤级错误定位重采样
StepBack: Step-Level Error-Localized Resampling for Efficient Test-Time Reasoning
Yang Ouyang · Jung-Eun Kim
中文摘要
自一致性（self-consistency, SC）是提升思维链（chain-of-thought）推理的常用解码策略，但需在预设预算下采样大量完整推理轨迹。其高效变体主要通过提前停止整条轨迹采样或在问题间分配采样数来降低成本。本文提出 StepBack，一种步骤级错误定位重采样方法，用于高效测试时推理。给定少量草稿轨迹，StepBack 利用轻量信号定位可能的首个错误位置，再对检测到的错误步骤及其后缀重采样，最后进行多数答案投票。在七个数学推理与多跳问答基准上，StepBack 平均达到与 SC 相当的准确率，同时将 token 用量最多降低 41.26%。此外，StepBack 本质上可与 Adaptive Consistency 等自适应推理方法结合，进一步提升其推理时效率。
关键要点
- 01自一致性需采样大量完整轨迹，现有高效变体仅做轨迹级或问题级预算分配
- 02StepBack 用轻量信号定位首个错误步骤，仅重采样该步骤及后缀
- 03七个数学推理与多跳问答基准上平均准确率与 SC 相当，token 用量最多降 41.26%
- 04可与 Adaptive Consistency 等自适应推理方法结合，进一步提升推理效率
解读
尚无解读。
原始英文摘要
arXiv:2611.08403v1 Announce Type: cross Abstract: Self-consistency (SC) is a widely used decoding strategy for improving chain-of-thought reasoning, but it requires sampling many complete reasoning trajectories with a preset budget. Its efficient variants reduce the cost mainly by stopping full-trajectory sampling early or distributing the number of samples across questions. In this paper, we propose StepBack, a step-level error-localized resampling method for efficient test-time reasoning. Given a small set of draft trajectories, StepBack localizes likely first error positions using lightweight signals and then resamples the detected error step and its suffix, followed by majority answer voting. Empirically, StepBack matches the accuracy of SC on average across seven mathematical reasoning and multi-hop QA benchmarks, while reducing token usage by up to 41.26%. In addition, by nature, StepBack can be combined with adaptive inference methods such as Adaptive Consistency, further improving their inference-time efficiency.