幻觉对于具有开放世界假设的LLM是不可避免的
Hallucination is Inevitable for LLMs with the Open World Assumption
AI总结：
本文论证在开放世界假设下LLM的幻觉不可避免，将其重新定义为泛化问题，并分类可纠正与不可避免情况，主张将其视为需容忍的结构性特征。
摘要：
大型语言模型（LLMs）展现出令人印象深刻的语言能力，但也产生不准确或捏造的输出，通常被称为“幻觉”。工程方法通常将幻觉视为需要最小化的缺陷，而形式化分析则论证了其理论上的不可避免性。然而，在考虑人工通用智能（AGI）所需条件时，这两种视角仍不完整。本文重新将“幻觉”框定为泛化问题的一种表现。在封闭世界假设下，即训练和测试分布一致时，幻觉可能得到缓解。然而，在开放世界假设下，即环境是无界的，幻觉变得不可避免。本文进一步对幻觉进行了分类，区分了在开放世界条件下可能被纠正的情况与看似不可避免的情况。在此基础上，本文提出“幻觉”不应仅被视为工程缺陷，而应被视为一种结构性特征，需要被容忍并与人类智能相兼容。
英文摘要：
Large Language Models (LLMs) exhibit impressive linguistic competence but also produce inaccurate or fabricated outputs, often called ``hallucinations''. Engineering approaches usually regard hallucination as a defect to be minimized, while formal analyses have argued for its theoretical inevitability. Yet both perspectives remain incomplete when considering the conditions required for artificial general intelligence (AGI). This paper reframes ``hallucination'' as a manifestation of the generalization problem. Under the Closed World assumption, where training and test distributions are consistent, hallucinations may be mitigated. Under the Open World assumption, however, where the environment is unbounded, hallucinations become inevitable. This paper further develops a classification of hallucination, distinguishing cases that may be corrected from those that appear unavoidable under open-world conditions. On this basis, it suggests that ``hallucination'' should be approached not merely as an engineering defect but as a structural feature to be tolerated and made compatible with human intelligence.
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。
