Measuring Meaning Composition in the Human Brain with Composition Scores from Large Language Models

arXiv:2403.04325v1 Announce Type: new
Abstract: The process of meaning composition, wherein smaller units like morphemes or words combine to form the meaning of phrases and sentences, is essential for human sentence comprehension. Despite extensive neurolinguistic research into the brain regions involved in meaning composition, a computational metric to quantify the extent of composition is still lacking. Drawing on the key-value memory interpretation of transformer feed-forward network blocks, we introduce the Composition Score, a novel model-based metric designed to quantify the …

abstract arxiv brain computational cs.ai cs.cl form human language language models large language large language models meaning measuring process research type units words