让多个问题一起推理，提升大模型准确率并降低计算成本
Batch-of-Thought: Cross-Instance Learning for Enhanced LLM Reasoning
- 将相关问题批量处理，通过对比发现优质推理模式
- 在6个基准上平均提升准确率，最高降低61%推理开销
- 适合需要高精度与低延迟的推理场景，无需额外训练
当前大语言模型推理系统独立处理每个查询，忽略了共享推理模式和一致性约束等跨实例信号。我们提出无需训练的批量思维（Batch-of-Thought, BoT）方法，通过联合处理相关查询实现跨实例学习。通过批次内比较分析，BoT能识别高质量推理模板、通过一致性检查发现错误，并分摊计算成本。我们在多智能体反思架构（BoT-R）中实现该方法，由反思器执行联合评估，挖掘孤立处理无法获得的互信息。在三种模型家族和六个基准上的实验表明，BoT-R持续提升准确率与置信度校准效果，同时推理成本最多降低61%。理论与实验分析揭示了批处理推理在何种情况下有效。代码已开源。
原文摘要 · Abstract (English)
Current Large Language Model reasoning systems process queries independently, discarding valuable cross-instance signals such as shared reasoning patterns and consistency constraints. We introduce Batch-of-Thought (BoT), a training-free method that processes related queries jointly to enable cross-instance learning. By performing comparative analysis across batches, BoT identifies high-quality reasoning templates, detects errors through consistency checks, and amortizes computational costs. We instantiate BoT within a multi-agent reflection architecture (BoT-R), where a Reflector performs joint evaluation to unlock mutual information gain unavailable in isolated processing. Experiments across three model families and six benchmarks demonstrate that BoT-R consistently improves accuracy and confidence calibration while reducing inference costs by up to 61%. Our theoretical and experimental analysis reveals when and why batch-aware reasoning benefits LLM systems. Our code is available at https://github.com/xuanyang19/BoT
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。