In a literate human brain, reading and writing are two doubly dissociable systems: a ventral decoding route impaired in pure alexia and a fronto-parietal encoding route impaired in pure agraphia, sharing a partial orthographic core. However, a decoder-only large language model (LLM) drives both from a single autoregressive path optimized for text, which is a recent cultural invention rather than an evolved instinct. We investigate how entangled this mechanism is by comparing an input-side 'reading code' $W_E$ with an output-side 'writing code' $W_U$, using an entanglement index $E \ ext{in} [0,1]$ calibrated against an independent-init floor and a tied ceiling. Across nine probes on GPT-2, OPT, Pythia (14M--1.4B), T5, and BERT/RoBERTa, two complementary levels agree in direction. In the weights, untied models hold one coupled but sub-ceiling code ($E=0.23$--$0.35$, far above floor) on a non-monotonic couple-then-differentiate trajectory, with $W_U$ drifting approximately $3.2 imes$ farther than $W_E$ in every frequency decile. Behaviorally, comprehension and production are positively coupled in all 12 non-degenerate models (sign test $p < 0.001$).
Blogger's Review: This paper reveals the underlying mechanisms of how large language models handle reading and writing tasks, providing deep insights into how these two cognitive functions are intertwined. This research may influence our fundamental understanding of language processing in future AI developments.