main concepts scientists sections
SimplePopularAdvanced

How a Neural Network Learns to Speak Like a Human express

Original: "Training-Driven Representational Geometry Modularization Predicts Brain Alignment in Language Models"
· Yixuan Liu, Zhiyuan Ma, Likai Tang, Runmin Gan, Xinche Zhang, Jinhao Li, Chao Xie, Sen Song
arXiv:2602.07539 · 2026-02-07 · CC BY 4.0 · 1 min
Neurons and Cognition Computation and Language
Large language models reorganize their layers as they age, starting to resemble the brain.

A language model is a multi-layered structure where each layer processes words in its own way. At the start of training, layers are chaotic, but gradually they organize into two teams. The first team works simply: its internal geometry smooths out, becoming less tangled. The second remains complex and intricate.

Layers with simple geometry turned out to be better predictors of activity in the brain's language areas. The simpler, the more accurate.

Scientists compared the model with fMRI data from people who listened to or read text. The similarity to the brain appeared unevenly: in temporal areas (responsible for the sound and meaning of words) it was fast and stable, while in frontal areas (responsible for building phrases) it came later and gradually. Moreover, for larger models, this connection is stronger.

Even when the model continued to learn, it was the simplicity of geometry that predicted brain similarity better than mere training experience.

It turns out that learning language is a path from chaos to order. The brain may do the same: smoothing some pathways of information processing while leaving others complex for flexibility.

Whole graph →

The layers of a language model can be compared to the floors of a building: lower floors break words into sounds and letters, upper floors build meaning from them. The study showed that 'simple' floors are closer to the entrance, and 'complex' ones are closer to the exit.

In the movie 'Her,' the artificial intelligence Samantha evolves and begins to understand people better than they understand themselves. This research is a step toward understanding how a machine could reach such a level.

Concepts
Large Language Models protein language model neural networks in spatial mapping generalization predictive learning neural networks predictive model training data regularization methods self-supervised learning training frame tumor segmentation transformers symplectic geometry medical image analysis deep learning in medical imaging neural network memory architecture medical image to genomic mapping learning rate hippocampus