Vectors from Larger Language Models Predict Human Reading Time and fMRI Data More Poorly when Dimensionality Expansion is Controlled

TL;DR

This study shows that larger language model vectors predict human reading times and fMRI data less accurately when controlling for vector dimensionality, indicating a misalignment that worsens with model size.

Contribution

It introduces a method to evaluate LLM scaling by controlling for predictor dimensionality, revealing inverse scaling effects in predicting human data.

Findings

Inverse scaling observed when controlling for predictor size.

Larger LLM vectors predict human data less accurately.

Misalignment between LLMs and human sentence processing increases with model size.

Abstract

The impressive linguistic abilities of large language models (LLMs) have recommended them as models of human sentence processing, with some conjecturing a positive 'quality-power' relationship (Wilcox et al., 2023), in which language models' (LMs') fit to psychometric data continues to improve as their ability to predict words in context increases. This is important because it suggests that elements of LLM architecture, such as veridical attention to context and a unique objective of predicting upcoming words, reflect the architecture of the human sentence processing faculty, and that any inadequacies in predicting human reading time and brain imaging data may be attributed to insufficient model complexity, which recedes as larger models become available. Recent studies (Oh and Schuler, 2023) have shown this scaling inverts after a point, as LMs become excessively large and accurate, when word prediction probability (as information-theoretic surprisal) is used as a predictor. Other studies propose the use of entire vectors from differently sized LLMs, still showing positive scaling (Schrimpf et al., 2021), casting doubt on the value of surprisal as a predictor, but do not control for the larger number of predictors in vectors from larger LMs. This study evaluates LLM scaling using entire LLM vectors, while controlling for the larger number of predictors in vectors from larger LLMs. Results show that inverse scaling obtains, suggesting that inadequacies in predicting human reading time and brain imaging data may be due to substantial misalignment between LLMs and human sentence processing, which worsens as larger models are used.