Language Models as Higher-Order Planning Formalizers
新论文探索利用语言模型实现高阶规划形式化,揭示AI推理与决策的前沿突破。
arXiv:2603.23844v2 Announce Type: replace Abstract: Recent work provides overwhelming evidence that LLMs, even those trained to scale their reasoning …
新论文探索利用语言模型实现高阶规划形式化,揭示AI推理与决策的前沿突破。
arXiv:2603.23844v2 Announce Type: replace Abstract: Recent work provides overwhelming evidence that LLMs, even those trained to scale their reasoning …
当大语言模型嵌入安全关键的推理流程,这个“叙事差距”可能成为致命短板,论文深入分析了LLM与求解器协作中的逻辑漏洞
arXiv:2606.19588v1 Announce Type: new Abstract: Formal tools such as SAT and SMT solvers are increasingly embedded in language model reasoning pipelin…
用停机问题检验LLM的极限:这篇论文通过理论分析和实验对比,揭示了大语言模型在程序终止性推理上的能力与局限
arXiv:2601.18987v5 Announce Type: replace-cross Abstract: Determining whether a program terminates is a central problem in computer science. Turing's …
美团开源LongCat-Flash-Prover,AI数学定理证明新SOTA,MiniF2F通过率97.1%
在常规的数学解题中,模型只需要“答对最终数值”即可,但数学定理证明不同,它要求极度严苛的逻辑链条,任何一句自然语言的模棱两可,都可能导致整个证明的崩塌。那么,如何让 AI 从“猜答案”走向“严谨证明”,成为复杂推理具有挑战的课题。为了解答这个问题,我们开源了专门用于数学形式化与定理证明的模型 —— …