离散难度是否足够？利用连续难度实现大语言模型中高效的自一致性
Is Discrete Difficulty Sufficient? Leveraging Continuous Difficulty for Efficient Self-Consistency in LLMs
浏览论文内容
中文总结 AI 辅助
本研究针对大语言模型自一致性（SC）token消耗高的问题，提出灵活自一致性（FSC），以连续信号估计难度动态调整路径数，在保持准确率的同时最高节省76% token。
中文摘要 AI 辅助
自一致性（Self-Consistency, SC）是一种解码策略，通过采样多样化推理路径并选择最一致的答案，在复杂推理问题上表现出强大性能。然而，生成多条推理路径带来的过多token消耗是SC的主要局限。为提升计算效率，多项研究提出了调整推理路径数量或根据问题难度差异化分配资源的策略。但现有多数方法将难度划分为少数固定级别，未能充分捕捉推理复杂度连续变化的特性。本研究提出灵活自一致性（Flexible Self-Consistency, FSC），将问题难度估计为连续信号并据此动态调整生成的推理路径数量。FSC使用预训练探测模型预测输入问题的输出熵，将其作为模型不确定性的指标，灵活控制采样预算。实验结果表明，在多种模型和基准测试中，FSC保持了与SC相当的准确率，同时实现了最高达76%的token节省。
英文摘要
Self-Consistency (SC) is a decoding strategy that samples diverse reasoning paths and selects the most consistent answer, demonstrating strong performance on complex reasoning problems. However, the excessive token consumption incurred by generating multiple reasoning paths has been identified as a major limitation of SC. To improve computational efficiency, several studies have proposed strategies that adjust the number of reasoning paths or allocate resources differentially according to problem difficulty. Nevertheless, most existing methods categorize difficulty into a few fixed levels, failing to fully capture the continuously varying nature of reasoning complexity. In this work, we propose Flexible Self-Consistency (FSC), which estimates problem difficulty as a continuous signal and dynamically adjusts the number of generated reasoning paths accordingly. FSC predicts the output entropy of an input question using a pre-trained probe and leverages it as an indicator of model uncertainty to flexibly control the sampling budget. Experimental results show that, across various models and benchmarks, FSC maintains accuracy comparable to SC while achieving token savings of up to 76%.
发表机构
- Konkuk University（建国大学）
- DATUMO INC.（DATUMO公司）
机构由 AI 辅助整理，请以论文原文为准。
