大型推理模型早停方法综述与效率可靠性实证审计
原标题：Early Stopping in Large Reasoning Models: A Survey and Empirical Audit of Efficiency and Reliability
AI 摘要
该论文对大型推理模型（LRM）的早停方法进行了综述与实证审计，在五个推理模型上覆盖数学、科学推理、代码生成和事实性任务，采用共享 rollout 重放协议评估。作者提出 Target-Integrated Token Efficiency（TITE）指标衡量相对 oracle 的效率，并发现后续位置的正确率可能低于解码噪声模型预测，强制终止后模型仍会继续推理，部分早停输出支撑推理较弱。研究还揭示了早停方法跨模型与任务迁移的泛化问题，代码已开源。
正文节选
Early Stopping in Large Reasoning Models: A Survey and Empirical Audit of Efficiency and Reliability Abstract Large reasoning models can improve accuracy by generating longer reasoning traces, but some of this computation may be unnecessary. Early-stopping methods aim to stop once enough reasoning is available to answer. Evaluating these methods requires accounting for the cost of deciding when to stop and the reliability of the resulting answer. We evaluate different methods across five reasoni