arXiv cs.AI· Quang Hung Nguyen, Hanh Nguyen, Thanh Tam Nguyen, Quoc Viet Hung Nguyen, Trinh Pham, Thanh Trung Huynh·· 1 天前AI 评分38
大型推理模型的早停方法：效率与可靠性综述及实证审计
Early Stopping in Large Reasoning Models: A Survey and Empirical Audit of Efficiency and Reliability
AI 导读
研究用共享 rollout 回放协议在 5 个推理模型上评测早停方法，覆盖数学与科学推理、代码生成和事实性任务，并提出 Target-Integrated Token Efficiency（TITE）以衡量相对 oracle 的 token 效率。实验发现，退出位置越靠后，正确率下降幅度超出拟合的解码噪声模型预测；可靠性检查还发现强制终止后推理仍在继续，部分早停正确的输出其支撑推理更弱。
正文 · 原文
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 reasoning models on mathematical and scientific reasoning, code generation, and factuality tasks by using a shared rollout replay protocol. We introduce Target-Integrated Token Efficiency (TITE) to measure efficiency relative to the oracle across accuracy targets. We also examine correctness at later exit positions, transfer across models and tasks, post-stop reasoning continuation, and retained justification quality. In our experiments, correctness can decrease at later positions beyond what the fitted decoding-noise model predicts. Reliability checks also reveal continued reasoning after forced termination and weaker supporting reasoning in some correct early-stopping outputs. The implementation can be found in this https URL.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.00166 [cs.AI] |
| (or arXiv:2611.00166v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2611.00166 arXiv-issued DOI via DataCite |
Submission history
From: Hung Nguyen Quang [view email]
[v1]
Fri, 30 Oct 2026 11:39:08 UTC (2,336 KB)
[v2]
Thu, 12 Nov 2026 13:12:13 UTC (2,263 KB)
来源：arXiv cs.AI · arxiv.org