大语言模型能感知时间吗？一项实证研究
Can LLMs Perceive Time? An Empirical Investigation
AI总结：
研究探讨大语言模型在时间估计上的局限性，通过68项任务和四类模型进行实验，发现模型在预任务估计、相对排序和事后回忆中均存在显著误差，表明模型缺乏对自身推理时间的体验性认知。
摘要：
大语言模型无法估计自身任务的持续时间。我们通过六十八项任务和四类模型家族的四个实验来调查这一限制。预任务估计比实际持续时间高4-7倍（p < 0.001），模型预测的人类分钟级时间用于完成秒级的任务。相对排序同样不理想：在设计暴露启发式依赖的任务对中，模型得分在或低于随机水平（GPT-5在反直觉对上得分为18%，p = 0.033），当复杂性标签误导时系统性失败。事后回忆与现实脱节——估计值在任一方向上与实际值相差一个数量级。这些失败在多步骤代理设置中持续存在，误差为5-10倍。模型在训练中具有关于持续时间的命题知识，但缺乏对自身推理时间的体验性基础，这对代理调度、规划和时间敏感场景有实际影响。
英文摘要：
Large language models cannot estimate how long their own tasks take. We investigate this limitation through four experiments across 68 tasks and four model families. Pre-task estimates overshoot actual duration by 4--7$\times$ ($p < 0.001$), with models predicting human-scale minutes for tasks completing in seconds. Relative ordering fares no better: on task pairs designed to expose heuristic reliance, models score at or below chance (GPT-5: 18\% on counter-intuitive pairs, $p = 0.033$), systematically failing when complexity labels mislead. Post-hoc recall is disconnected from reality -- estimates diverge from actuals by an order of magnitude in either direction. These failures persist in multi-step agentic settings, with errors of 5--10$\times$. The models possess propositional knowledge about duration from training but lack experiential grounding in their own inference time, with practical implications for agent scheduling, planning and time-critical scenarios.
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。
