大语言模型在滑动拼图中先天规划能力的可验证效率极限
Verifiable Efficiency Limits of Innate LLM Planning in Sliding Puzzles
AI 导读
一项 arXiv 研究提出可验证框架 SPE=R×Q×E，通过 4×4 滑动拼图上 8,400 份完整规划输出、覆盖七种模型条件，剥离工具与执行反馈来衡量大语言模型的先天规划能力。结果显示，即便能完成全部试次的模型，在更长步数视野下行动效率也会下降；SPE 能区分规划协调失败与完成低效，使无辅助规划的长视野局限可被定量验证。
正文 · 原文
Abstract:Tool-assisted planning success can reflect external search and verification, leaving a model's unaided planning ability unresolved. We isolate innate planning through complete plans generated in one response without tools or execution feedback. We propose a verifiable framework that decomposes success-weighted plan efficiency as $\mathrm{SPE}=R\times Q\times E$: execution reliability, planning coordination among legal plans, and action economy among completed plans. Trajectory replay and certified optima verify these factors; the joint score includes unsuccessful attempts. We evaluate 8,400 full-plan outputs from seven model conditions on $4\times4$ sliding puzzles, controlling exact horizon and obstacle count in the primary cohort. Models that complete every trial still become less economical at longer horizons, while higher efficiency among completed plans can coexist with poor goal attainment. SPE separates the overall capability hierarchy, while its factors distinguish coordination failure from inefficient completion. The framework makes long-horizon limitations in unaided planning quantitatively verifiable.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.36415 [cs.AI] |
| (or arXiv:2610.36415v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.36415 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yezhou Cheng [view email]
[v1]
Mon, 26 Oct 2026 23:53:23 UTC (112 KB)
来源：arXiv cs.AI · arxiv.org