Capturing AI's Attention: Physics of Repetition, Hallucination, Bias and Beyond

作者：F. Huo, Neil F. Johnson · 发表于：arXiv.org · 年份：2025 · DOI：10.48550/arXiv.2504.04600 · 被引用次数：5 · 研究领域：Computer Science、Physics、Mathematics

We derive a first-principles physics theory of the AI engine at the heart of LLMs' 'magic' (e.g. ChatGPT, Claude): the basic Attention head. The theory allows a quantitative analysis of outstanding AI challenges such as output repetition, hallucination and harmful content, and bias (e.g. from training and fine-tuning). Its predictions are consistent with large-scale LLM outputs. Its 2-body form suggests why LLMs work so well, but hints that a generalized 3-body Attention would make such AI work even better. Its similarity to a spin-bath means that existing Physics expertise could immediately be harnessed to help Society ensure AI is trustworthy and resilient to manipulation.