揭示大模型幻觉的不可避免性，并提出两种破解路径。
Hallucination as a Computational Boundary: A Hierarchy of Inevitability and the Oracle Escape
- 将大模型视为概率图灵机，构建计算必然性层级
- 证明幻觉在对角化等边界下不可避免，但可借外接知识或持续学习突破
- 为RAG和持续学习提供形式化理论，适合安全可信AI研究者
大语言模型的幻觉现象是其可靠部署的核心障碍。本文通过构建‘计算必然性层级’，将大语言模型形式化为概率图灵机，首次证明幻觉在对角化、不可计算性和信息论边界下是不可避免的，依据新提出的‘学习者泵引理’。然而，本文提出两条‘逃逸路径’：一是将检索增强生成（RAG）建模为预言机，证明其可通过‘计算跃迁’实现绝对逃逸，首次为RAG的有效性提供形式化理论；二是将持续学习形式化为‘内化预言机’机制，并通过新型神经博弈论框架实现该路径。最后，本文提出人工智能安全的新原则——计算类对齐（CCA），要求任务复杂度与系统实际算力严格匹配，为人工智能的安全应用提供理论支撑。
原文摘要 · Abstract (English)
The illusion phenomenon of large language models (LLMs) is the core obstacle to their reliable deployment. This article formalizes the large language model as a probabilistic Turing machine by constructing a "computational necessity hierarchy", and for the first time proves the illusions are inevitable on diagonalization, incomputability, and information theory boundaries supported by the new "learner pump lemma". However, we propose two "escape routes": one is to model Retrieval Enhanced Generations (RAGs) as oracle machines, proving their absolute escape through "computational jumps", providing the first formal theory for the effectiveness of RAGs; The second is to formalize continuous learning as an "internalized oracle" mechanism and implement this path through a novel neural game theory framework. Finally, this article proposes a feasible new principle for artificial intelligence security - Computational Class Alignment (CCA), which requires strict matching between task complexity and the actual computing power of the system, providing theoretical support for the secure application of artificial intelligence.
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。