Language models outperform cloze predictability in a cognitive model of reading.

Although word predictability is commonly considered an important factor in reading, sophisticated accounts of predictability in theories of reading are lacking. Computational models of reading traditionally use cloze norming as a proxy of word predictability, but what cloze norms precisely capture r...

Full description

Saved in:
Bibliographic Details
Main Authors: Adrielli Tina Lopes Rego, Joshua Snell, Martijn Meeter
Format: Article
Language:English
Published: Public Library of Science (PLoS) 2024-09-01
Series:PLoS Computational Biology
Online Access:https://doi.org/10.1371/journal.pcbi.1012117
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850250843651047424
author Adrielli Tina Lopes Rego
Joshua Snell
Martijn Meeter
author_facet Adrielli Tina Lopes Rego
Joshua Snell
Martijn Meeter
author_sort Adrielli Tina Lopes Rego
collection DOAJ
description Although word predictability is commonly considered an important factor in reading, sophisticated accounts of predictability in theories of reading are lacking. Computational models of reading traditionally use cloze norming as a proxy of word predictability, but what cloze norms precisely capture remains unclear. This study investigates whether large language models (LLMs) can fill this gap. Contextual predictions are implemented via a novel parallel-graded mechanism, where all predicted words at a given position are pre-activated as a function of contextual certainty, which varies dynamically as text processing unfolds. Through reading simulations with OB1-reader, a cognitive model of word recognition and eye-movement control in reading, we compare the model's fit to eye-movement data when using predictability values derived from a cloze task against those derived from LLMs (GPT-2 and LLaMA). Root Mean Square Error between simulated and human eye movements indicates that LLM predictability provides a better fit than cloze. This is the first study to use LLMs to augment a cognitive model of reading with higher-order language processing while proposing a mechanism on the interplay between word predictability and eye movements.
format Article
id doaj-art-fe3098f9dfd2402088a5afd7c1ec4437
institution OA Journals
issn 1553-734X
1553-7358
language English
publishDate 2024-09-01
publisher Public Library of Science (PLoS)
record_format Article
series PLoS Computational Biology
spelling doaj-art-fe3098f9dfd2402088a5afd7c1ec44372025-08-20T01:58:04ZengPublic Library of Science (PLoS)PLoS Computational Biology1553-734X1553-73582024-09-01209e101211710.1371/journal.pcbi.1012117Language models outperform cloze predictability in a cognitive model of reading.Adrielli Tina Lopes RegoJoshua SnellMartijn MeeterAlthough word predictability is commonly considered an important factor in reading, sophisticated accounts of predictability in theories of reading are lacking. Computational models of reading traditionally use cloze norming as a proxy of word predictability, but what cloze norms precisely capture remains unclear. This study investigates whether large language models (LLMs) can fill this gap. Contextual predictions are implemented via a novel parallel-graded mechanism, where all predicted words at a given position are pre-activated as a function of contextual certainty, which varies dynamically as text processing unfolds. Through reading simulations with OB1-reader, a cognitive model of word recognition and eye-movement control in reading, we compare the model's fit to eye-movement data when using predictability values derived from a cloze task against those derived from LLMs (GPT-2 and LLaMA). Root Mean Square Error between simulated and human eye movements indicates that LLM predictability provides a better fit than cloze. This is the first study to use LLMs to augment a cognitive model of reading with higher-order language processing while proposing a mechanism on the interplay between word predictability and eye movements.https://doi.org/10.1371/journal.pcbi.1012117
spellingShingle Adrielli Tina Lopes Rego
Joshua Snell
Martijn Meeter
Language models outperform cloze predictability in a cognitive model of reading.
PLoS Computational Biology
title Language models outperform cloze predictability in a cognitive model of reading.
title_full Language models outperform cloze predictability in a cognitive model of reading.
title_fullStr Language models outperform cloze predictability in a cognitive model of reading.
title_full_unstemmed Language models outperform cloze predictability in a cognitive model of reading.
title_short Language models outperform cloze predictability in a cognitive model of reading.
title_sort language models outperform cloze predictability in a cognitive model of reading
url https://doi.org/10.1371/journal.pcbi.1012117
work_keys_str_mv AT adriellitinalopesrego languagemodelsoutperformclozepredictabilityinacognitivemodelofreading
AT joshuasnell languagemodelsoutperformclozepredictabilityinacognitivemodelofreading
AT martijnmeeter languagemodelsoutperformclozepredictabilityinacognitivemodelofreading