一篇新的研究论文提出，大型语言模型（LLM）是世界模型的特化形式，而非独立的类别。该论文认为，预测 token 的 LLM 可以被视为模拟现实的世界模型的退化案例。它表明，现有的 LLM 架构与更先进的世界模型之间存在一个连续的光谱，并且研究中已在探索潜在的中间步骤。 AI
影响 这项研究重新构建了对 LLM 的理解，提出了一个与世界模型统一的理论框架，并可能指导未来的架构发展。
排序理由 该集群包含一篇讨论 LLM 和世界模型理论方面的研究论文。
在 Hugging Face Daily Papers 阅读 →
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2606.28127v1 Announce Type: cross Abstract: The AI community has framed the relationship between large language models (LLMs) and world models as a dichotomy: LLMs predict tokens; world models simulate reality. Yann LeCun argues in 2022 that reaching general intelligence re…
The AI community has framed the relationship between large language models (LLMs) and world models as a dichotomy: LLMs predict tokens; world models simulate reality. Yann LeCun argues in 2022 that reaching general intelligence requires abandoning autoregressive token prediction …
The AI community has framed the relationship between large language models (LLMs) and world models as a dichotomy: LLMs predict tokens; world models simulate reality. Yann LeCun argues in 2022 that reaching general intelligence requires abandoning autoregressive token prediction …