@inproceedings{nair-resnik-2023-words,
title = "Words, Subwords, and Morphemes: What Really Matters in the Surprisal-Reading Time Relationship?",
author = "Nair, Sathvik and
Resnik, Philip",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.752",
doi = "10.18653/v1/2023.findings-emnlp.752",
pages = "11251--11260",
abstract = "An important assumption that comes with using LLMs on psycholinguistic data has gone unverified. LLM-based predictions are based on subword tokenization, not decomposition of words into morphemes. Does that matter? We carefully test this by comparing surprisal estimates using orthographic, morphological, and BPE tokenization against reading time data. Our results replicate previous findings and provide evidence that *in the aggregate*, predictions using BPE tokenization do not suffer relative to morphological and orthographic segmentation. However, a finer-grained analysis points to potential issues with relying on BPE-based tokenization, as well as providing promising results involving morphologically-aware surprisal estimates and suggesting a new method for evaluating morphological prediction.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="nair-resnik-2023-words">
<titleInfo>
<title>Words, Subwords, and Morphemes: What Really Matters in the Surprisal-Reading Time Relationship?</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sathvik</namePart>
<namePart type="family">Nair</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Philip</namePart>
<namePart type="family">Resnik</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-12</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2023</title>
</titleInfo>
<name type="personal">
<namePart type="given">Houda</namePart>
<namePart type="family">Bouamor</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Juan</namePart>
<namePart type="family">Pino</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kalika</namePart>
<namePart type="family">Bali</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Singapore</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>An important assumption that comes with using LLMs on psycholinguistic data has gone unverified. LLM-based predictions are based on subword tokenization, not decomposition of words into morphemes. Does that matter? We carefully test this by comparing surprisal estimates using orthographic, morphological, and BPE tokenization against reading time data. Our results replicate previous findings and provide evidence that *in the aggregate*, predictions using BPE tokenization do not suffer relative to morphological and orthographic segmentation. However, a finer-grained analysis points to potential issues with relying on BPE-based tokenization, as well as providing promising results involving morphologically-aware surprisal estimates and suggesting a new method for evaluating morphological prediction.</abstract>
<identifier type="citekey">nair-resnik-2023-words</identifier>
<identifier type="doi">10.18653/v1/2023.findings-emnlp.752</identifier>
<location>
<url>https://aclanthology.org/2023.findings-emnlp.752</url>
</location>
<part>
<date>2023-12</date>
<extent unit="page">
<start>11251</start>
<end>11260</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Words, Subwords, and Morphemes: What Really Matters in the Surprisal-Reading Time Relationship?
%A Nair, Sathvik
%A Resnik, Philip
%Y Bouamor, Houda
%Y Pino, Juan
%Y Bali, Kalika
%S Findings of the Association for Computational Linguistics: EMNLP 2023
%D 2023
%8 December
%I Association for Computational Linguistics
%C Singapore
%F nair-resnik-2023-words
%X An important assumption that comes with using LLMs on psycholinguistic data has gone unverified. LLM-based predictions are based on subword tokenization, not decomposition of words into morphemes. Does that matter? We carefully test this by comparing surprisal estimates using orthographic, morphological, and BPE tokenization against reading time data. Our results replicate previous findings and provide evidence that *in the aggregate*, predictions using BPE tokenization do not suffer relative to morphological and orthographic segmentation. However, a finer-grained analysis points to potential issues with relying on BPE-based tokenization, as well as providing promising results involving morphologically-aware surprisal estimates and suggesting a new method for evaluating morphological prediction.
%R 10.18653/v1/2023.findings-emnlp.752
%U https://aclanthology.org/2023.findings-emnlp.752
%U https://doi.org/10.18653/v1/2023.findings-emnlp.752
%P 11251-11260
Markdown (Informal)
[Words, Subwords, and Morphemes: What Really Matters in the Surprisal-Reading Time Relationship?](https://aclanthology.org/2023.findings-emnlp.752) (Nair & Resnik, Findings 2023)
ACL