大语言模型中的早期停止思维链
Early Stopping Chain-of-thoughts in Large Language Models
AI总结：
本文提出了一种在推理阶段减少思维链生成长度的方法ES-CoT，通过检测答案收敛并提前停止来降低推理成本，同时保持与标准思维链相当的准确性。
摘要：
大语言模型（LLMs）在通过生成长链式思维（CoT）解决复杂问题时表现出卓越的能力，但这种长CoT会带来较高的推理成本。先前的推理阶段高效推理方法要么需要白盒模型监控推理过程，要么通过直接提示不可靠。为此，我们引入了ES-CoT，一种在推理时缩短CoT生成的方法，通过检测答案收敛并提前停止几乎不损失性能。当观察到推理过程中的语言标记（如“wait”）时，我们提示LLM输出其当前最终答案，称为步骤答案。我们跟踪连续相同步骤答案的运行长度作为答案收敛的度量。我们通过实证和理论证明，步骤答案稳定地收敛到最终答案，且大运行长度跳跃可靠地标记这种收敛。在六个跨三个LLM的推理数据集上的实验表明，ES-CoT在平均上减少了16.08%的推理令牌数量，同时保持与标准CoT相当的准确性。
英文摘要：
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.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
