让大模型在推理中途提前停止，节省算力还更准。
TERMINATOR: Learning Optimal Exit Points for Early Stopping in Chain-of-Thought Reasoning
- 根据首次出现答案的位置，训练模型预测最优推理长度。
- 在4个数据集上平均减少14%-55%的推理长度。
- 适合需要高效推理的大模型应用，如数学解题、代码生成。
大型推理模型（LRMs）通过思维链（CoT）推理在复杂任务中表现优异，但常因过度思考而浪费计算资源。已有研究发现存在最优推理长度，此时截断可显著缩短输出且性能几乎不变。然而，实际任务中该长度高度依赖于具体任务与模型，难以确定。本文提出 Terminator，一种基于推理阶段的早停策略，核心思想是首次生成最终答案的位置具有可预测性，利用这些位置构建最优推理长度数据集以训练终止模型。实验表明，Terminator 在 MATH-500、AIME 2025、HumanEval、GPQA 四个数据集上平均减少 14%-55% 的 CoT 长度，优于现有方法，推理延迟降低超过 2 倍。
原文摘要 · Abstract (English)
Large Reasoning Models (LRMs) achieve impressive performance on complex reasoning tasks via Chain-of-Thought (CoT) reasoning, which enables them to generate intermediate thinking tokens before arriving at the final answer. However, LRMs often suffer from significant overthinking, spending excessive compute time even after the answer is generated early on. Prior work has identified the existence of an optimal reasoning length such that truncating reasoning at this point significantly shortens CoT outputs with virtually no change in performance. However, determining optimal CoT lengths for practical datasets is highly non-trivial as they are fully task and model-dependent. In this paper, we precisely address this and design Terminator, an early-exit strategy for LRMs at inference to mitigate overthinking. The central idea underpinning Terminator is that the first arrival of an LRM's final answer is often predictable, and we leverage these first answer positions to create a novel dataset of optimal reasoning lengths to train Terminator. Powered by this approach, Terminator achieves significant reductions in CoT lengths of 14%-55% on average across four challenging practical datasets: MATH-500, AIME 2025, HumanEval, and GPQA, while outperforming current state-of-the-art methods and reducing inference latency by more than 2x compared to the original LRM.
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。