AI 中文总结
研究探讨大语言模型中阅读和写作代码的纠缠情况，通过纠缠指数比较输入输出代码，发现权重上未绑定模型有耦合代码，行为上理解与生成正耦合，此耦合普遍，贡献是量化及跨层一致性，定位LLMs为独特思维点。
AI 中文摘要
在有读写能力的人类大脑中，阅读和写作是两个相互独立的系统：腹侧解码路径（在纯失读症中受损）和额顶叶编码路径（在纯失写症中受损），共享部分正字法核心。而仅解码器的大语言模型（LLM）则通过在文本上优化的单个自回归路径驱动这两者，这是一种文化发明而非进化本能。我们通过纠缠指数\(E\in[0,1]\)（CKA、Procrustes残差、互\(k\)-NN）来比较输入侧的“阅读代码”\(W_E\)和输出侧的“写作代码”\(W_U\)，该指数是根据独立初始化下限和绑定上限校准的。在对GPT-2、OPT、Pythia（14M - 1.4B）、T5以及BERT/RoBERTa的九个探针测试中（六个巩固现有结果，三个引入读写分析），两个互补层面在方向上一致。在权重方面，未绑定模型在非单调的耦合然后分化轨迹上有一个耦合但低于上限的代码（\(E = 0.23 - 0.35\)，远高于下限），\(W_U\)在每个频率十分位数上的漂移比\(W_E\)远约\(3.2\)倍。在行为上，所有12个非退化模型中的理解和生成呈正耦合（符号检验\(p < 0.001\)），与大脑的双重分离相反。这种耦合是普遍的，不仅限于解码器：编码器 - 解码器在表示上分离了这两个路径（高达\(0.96\)），但在行为上仍保持耦合。我们明确报告了我们的无效结果（几何结构到行为的桥梁是无效的，\(\rho = 0.00\)）。由于单个前向路径使得某种耦合在理论上是可预期的，我们的贡献在于对其进行量化以及跨层面的一致性；通过类比而非同源性，这将大语言模型定位为可能思维空间中的一个独特点。
英文摘要
In the literate human brain, reading and writing doubly dissociate: a ventral decoding route (pure alexia) and a fronto-parietal encoding route (pure agraphia), sharing a partial orthographic core. A decoder-only large language model (LLM) drives both from one autoregressive path optimized on text (a \emph{cultural} invention, not an evolved instinct). We ask how entangled it is, comparing an input-side ``reading code'' $\mathbf{W}_{E}$ with an output-side ``writing code'' $\mathbf{W}_{U}$ via an index $\mathcal{E}\in[0,1]$ (CKA, Procrustes residual, mutual $k$-NN) calibrated against an independent-init floor and tied ceiling. On GPT-2, OPT and Pythia (14M--1.4B), untied models hold one \emph{coupled but sub-ceiling} code ($\mathcal{E}=0.23$--$0.35$, far above floor) on a non-monotonic couple-then-differentiate trajectory, $\mathbf{W}_{U}$ drifting $\sim$3.2$\times$ farther than $\mathbf{W}_{E}$ in every decile. Equally informative is a negative: the matching behavioural test, that comprehension and production fail together rather than dissociate, cannot be run. For minimal pairs the alexia analogue is empty by theorem: greedy production implies a vocabulary-wide argmax, so it wins the pairwise ranking. Differential-damage indices are not scale-identified: heavy-tailed damage makes linear standardizations collapse onto their larger term, and the rank transform fixing this is bounded, so its null saturates. Both scores also contain the target's log-probability, which alone explains most of their variance and manufactures the apparent coupling. We withdraw a coupling statistic, a cross-level bridge and a separation measure. In a model reading and writing off one next-token distribution, no output-side pair isolates either ability: entanglement needing no index to see. By analogy, not homology, this situates LLMs in the space of possible minds.
