Rui He

Activation-Guided Neuron Intervention to Induce Alzheimer's-Related Computational Language Phenotypes in a Large Language Model

It is shown that neurons identified solely from clinical language differences can influence behavior across multiple cognitive domains, providing proof of concept for an AD-related computational phenotype and a controlled framework for experimentally examining links between language and broader cognitive dysfunction.

Large language model linguistic perplexity in childhood onset psychosis: unique features and developmental trends

Psychosis involves a deficit of information compression in connected speech

Large language models (LLMs) with human-like performance on linguistic tasks have transformed the study of language in neurodiverse conditions. LLMs provide representations of linguistic input in the form of high-dimensional vectors (embeddings), and next-token predictions computed from these embeddings. Previous cross...

Structural brain complexity is associated with linguistic complexity in psychosis

Abstract Background Brain-structural and language abnormalities in people with psychotic symptoms (PSx) have long been reported. A key question is whether language alterations are a potential behavioral readout of brain cortical gray matter changes. Adding to previous evidence for this in the cases of cortical thinning...

Cross-lingual brain-language model alignment is robust but challenges hierarchical and computational accounts

The results suggest that brain-LLM alignment is more robust across languages, transformer depth, and model architectures than previously appreciated, but less informative about shared computational mechanisms.

The Changing Geometry of Grammar: Dimensionality and Neighborhood Reorganization across Transformer Layers

This work investigates the layer-wise evolution of ID, and shows that closed-class items expand earlier and collapse sooner than open-class ones, and recovers a token's grammatical role, and uses them to interpret how the semantic content of each PoS evolves across layers in a downstream classification task.

Distinct dynamics of conceptual and referential disruptions in human reading and large language model processing

Linguistic meaning is grounded in conceptual content, from which reference to particular entities emerges as words enter discourse. To examine the processing dynamics associated with these two dimensions of meaning, we selectively disrupted conceptual or referential information in short narratives and traced the result...