Exploring Similarity between Neural and LLM Trajectories in Language Processing

2025-09-29

AI Summary

This study investigates representational similarities between large language models (LLMs) and the human brain during language processing. We propose a dynamic neuro-model trajectory alignment framework that integrates ridge regression with multidimensional trajectory features—magnitude, angular dynamics, and uncertainty—to systematically compare hidden-state representations from 16 multilingual LLMs against EEG-derived neural responses. Our analysis reveals, for the first time, a significant correspondence between high-layer LLM representations and the N400 event-related potential, identifying key network layers underlying semantic integration. Crucially, we uncover a fundamental mechanistic divergence: human cortical processing relies on continuous, iterative computation, whereas LLMs exhibit discrete, stepwise activation patterns. These findings replicate robustly across English and Chinese language tasks. The work establishes a novel interdisciplinary paradigm bridging AI interpretability and cognitive neuroscience, providing empirically grounded insights into the neural plausibility of language model architectures.

Technology Category

• Natural Language Processing: (Large) Language Models
• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Understanding the similarity between large language models (LLMs) and human brain activity is crucial for advancing both AI and cognitive neuroscience. In this study, we provide a multilinguistic, large-scale assessment of this similarity by systematically comparing 16 publicly available pretrained LLMs with human brain responses during natural language processing tasks in both English and Chinese. Specifically, we use ridge regression to assess the representational similarity between LLM embeddings and electroencephalography (EEG) signals, and analyze the similarity between the "neural trajectory" and the "LLM latent trajectory." This method captures key dynamic patterns, such as magnitude, angle, uncertainty, and confidence. Our findings highlight both similarities and crucial differences in processing strategies: (1) We show that middle-to-high layers of LLMs are central to semantic integration and correspond to the N400 component observed in EEG; (2) The brain exhibits continuous and iterative processing during reading, whereas LLMs often show discrete, stage-end bursts of activity, which suggests a stark contrast in their real-time semantic processing dynamics. This study could offer new insights into LLMs and neural processing, and also establish a critical framework for future investigations into the alignment between artificial intelligence and biological intelligence.

Problem

• Compares neural trajectories and LLM latent dynamics in language processing
• Analyzes EEG-LLM similarity across 16 models and two languages
• Identifies differences in continuous brain vs discrete LLM processing

Innovation

• Used ridge regression for neural-LLM similarity analysis
• Compared neural and LLM trajectories across multiple languages
• Analyzed dynamic patterns like magnitude, angle and uncertainty