只在不确定环节分支，提升大模型推理效率与准确率
Entropy-Gated Branching for Efficient Test-Time Reasoning
- 仅在高不确定性步骤分支，减少无效计算
- 数学和金融任务准确率提升22.6%，速度比束搜索快31%-75%
- 适合需要高效推理的部署场景，如实时问答系统
测试时计算方法可显著提升大语言模型的推理能力与解题准确率，但需大量计算资源，多数计算浪费在低多样性、高置信度的分支上。我们发现少数不确定的推理步骤对最终预测准确率影响极大，于是在这些关键节点进行分支能生成更多样、更高质量的候选推理路径。为此提出熵门控分支（EGB），仅在高不确定性步骤分支，并用轻量级验证器剪枝。在数学与金融推理基准上，EGB相比标准推理准确率提升22.6%，且在数学任务上运行速度比测试时束搜索快31%至75%，同时性能更高。结果表明，推理过程中的动态资源分配可大幅提升效率与效果，为增强大模型推理能力提供更可持续的路径。
原文摘要 · Abstract (English)
Test-time compute methods can significantly improve the reasoning capabilities and problem-solving accuracy of large language models (LLMs). However, these approaches require substantially more computational resources, with most compute wasted on exploring low-diversity branches where the model already exhibits high confidence. We observe that a small subset of uncertain reasoning steps has a disproportionately large impact on final prediction accuracy, and branching at these critical junctures tends to yield more diverse and higher-quality candidate reasoning steps. We propose Entropy-Gated Branching (EGB), which branches only at high-uncertainty steps and prunes expansions with a lightweight verifier. On mathematical and financial reasoning benchmarks, EGB improves accuracy by 22.6% over standard inference while operating 31%-75% faster across math benchmarks than test-time beam search with higher performance. Our results show that dynamic resource allocation during inference can substantially improve both efficiency and effectiveness, offering a more scalable pathway to enhanced LLM reasoning capabilities.
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。