通过检测思维链收敛提前停止，显著降低大模型推理成本。
Early Stopping Chain-of-thoughts in Large Language Models
- 用语言标记触发输出中间答案，连续相同答案长度判断收敛
- 平均减少16.08%推理词元数，准确率与原方法基本持平
- 无需白盒访问，适用于任意大模型，适合部署优化
推理型大语言模型通过生成长思维链（CoT）解决复杂问题，但带来高推理开销。现有高效推理方法或需白盒模型监控，或通过直接提示不可靠。为此，我们提出ES-CoT，一种推理阶段的高效方法：通过检测思维链中语言标记（如“wait”）触发输出当前最终答案（步答案），并追踪连续相同步答案的长度作为收敛指标。实证与理论均表明，步答案会稳定收敛至最终答案，且长跳变可靠标志收敛。在三个大模型上的六组推理数据集实验显示，ES-CoT平均减少16.08%的推理词元，同时保持与标准CoT相当的准确率。
原文摘要 · Abstract (English)
Reasoning large language models (LLMs) have demonstrated superior capacities in solving complicated problems by generating long chain-of-thoughts (CoT), but such a lengthy CoT incurs high inference costs. Previous methods on inference-stage efficient reasoning either require white-box models to monitor the reasoning process or are not reliable through direct prompting. In response, we introduce ES-CoT, an inference-time method that shortens CoT generation by detecting answer convergence and stopping early with almost no performance loss. When observing a linguistic marker (such as "wait") in the reasoning process, we prompt the LLM to output its current final answer, denoted as a step answer. We then track the run length of consecutive identical step answers as a measure of answer convergence. We show both empirically and theoretically that step answers steadily converge to the final answer, and large run-length jumps reliably mark this convergence. Experiments on six reasoning datasets across three LLMs show that ES-CoT reduces the number of inference tokens by 16.08% on average while maintaining accuracy comparable to standard CoT.
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。