@inproceedings{ding-etal-2026-reading,
title = "Reading Acquisition Order out of {B}aby{LM} Representations: Child-Scale Training Produces Developmentally Aligned Semantic Spaces",
author = "Ding, Chenyi and
Houghton, Zachary Nicholas and
Jasbi, Masoud",
editor = "Choshen, Leshem and
Cotterell, Ryan and
Gul, Mustafa Omer and
Jumelet, Jaap and
Linzen, Tal and
Mueller, Aaron and
Salhan, Suchir and
Shah, Raj Sanjay and
Warstadt, Alex and
Wilcox, Ethan Gotlieb",
booktitle = "Proceedings of the Second {B}aby{LM} Workshop",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.babylm-main.20/",
pages = "250--259",
ISBN = "979-8-89176-475-0",
abstract = "Children learn words in a strikingly consistent order, and much of this order can be predicted from word-level properties such as frequency and concreteness. We ask whether acquisition order information is also implicitly present in the internal representations of language models, and what linguistic factors drive this. We probe 18 public checkpoints that span three architecture families at matched size, data scales from 10 million to roughly 180 billion words, and the winning systems and official baselines of the three previous BabyLM challenges. A ridge read-out of frozen hidden states predicts Wordbank age of acquisition for 619 CDI words, with the probing layer chosen on training folds only. Models trained on child-scale data match or exceed their full-scale counterparts in every family. Masked and hybrid training objectives yield more aligned spaces than purely causal training, and challenge winners are not the most aligned systems. Geometric measures of semantic structure add predictive value beyond classic predictors only for child-scale models. AoA decodability thus provides a training-free measure of developmental alignment for BabyLM evaluation."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ding-etal-2026-reading">
<titleInfo>
<title>Reading Acquisition Order out of BabyLM Representations: Child-Scale Training Produces Developmentally Aligned Semantic Spaces</title>
</titleInfo>
<name type="personal">
<namePart type="given">Chenyi</namePart>
<namePart type="family">Ding</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zachary</namePart>
<namePart type="given">Nicholas</namePart>
<namePart type="family">Houghton</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Masoud</namePart>
<namePart type="family">Jasbi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Second BabyLM Workshop</title>
</titleInfo>
<name type="personal">
<namePart type="given">Leshem</namePart>
<namePart type="family">Choshen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ryan</namePart>
<namePart type="family">Cotterell</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mustafa</namePart>
<namePart type="given">Omer</namePart>
<namePart type="family">Gul</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jaap</namePart>
<namePart type="family">Jumelet</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tal</namePart>
<namePart type="family">Linzen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aaron</namePart>
<namePart type="family">Mueller</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Suchir</namePart>
<namePart type="family">Salhan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Raj</namePart>
<namePart type="given">Sanjay</namePart>
<namePart type="family">Shah</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alex</namePart>
<namePart type="family">Warstadt</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ethan</namePart>
<namePart type="given">Gotlieb</namePart>
<namePart type="family">Wilcox</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-475-0</identifier>
</relatedItem>
<abstract>Children learn words in a strikingly consistent order, and much of this order can be predicted from word-level properties such as frequency and concreteness. We ask whether acquisition order information is also implicitly present in the internal representations of language models, and what linguistic factors drive this. We probe 18 public checkpoints that span three architecture families at matched size, data scales from 10 million to roughly 180 billion words, and the winning systems and official baselines of the three previous BabyLM challenges. A ridge read-out of frozen hidden states predicts Wordbank age of acquisition for 619 CDI words, with the probing layer chosen on training folds only. Models trained on child-scale data match or exceed their full-scale counterparts in every family. Masked and hybrid training objectives yield more aligned spaces than purely causal training, and challenge winners are not the most aligned systems. Geometric measures of semantic structure add predictive value beyond classic predictors only for child-scale models. AoA decodability thus provides a training-free measure of developmental alignment for BabyLM evaluation.</abstract>
<identifier type="citekey">ding-etal-2026-reading</identifier>
<location>
<url>https://aclanthology.org/2026.babylm-main.20/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>250</start>
<end>259</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Reading Acquisition Order out of BabyLM Representations: Child-Scale Training Produces Developmentally Aligned Semantic Spaces
%A Ding, Chenyi
%A Houghton, Zachary Nicholas
%A Jasbi, Masoud
%Y Choshen, Leshem
%Y Cotterell, Ryan
%Y Gul, Mustafa Omer
%Y Jumelet, Jaap
%Y Linzen, Tal
%Y Mueller, Aaron
%Y Salhan, Suchir
%Y Shah, Raj Sanjay
%Y Warstadt, Alex
%Y Wilcox, Ethan Gotlieb
%S Proceedings of the Second BabyLM Workshop
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-475-0
%F ding-etal-2026-reading
%X Children learn words in a strikingly consistent order, and much of this order can be predicted from word-level properties such as frequency and concreteness. We ask whether acquisition order information is also implicitly present in the internal representations of language models, and what linguistic factors drive this. We probe 18 public checkpoints that span three architecture families at matched size, data scales from 10 million to roughly 180 billion words, and the winning systems and official baselines of the three previous BabyLM challenges. A ridge read-out of frozen hidden states predicts Wordbank age of acquisition for 619 CDI words, with the probing layer chosen on training folds only. Models trained on child-scale data match or exceed their full-scale counterparts in every family. Masked and hybrid training objectives yield more aligned spaces than purely causal training, and challenge winners are not the most aligned systems. Geometric measures of semantic structure add predictive value beyond classic predictors only for child-scale models. AoA decodability thus provides a training-free measure of developmental alignment for BabyLM evaluation.
%U https://aclanthology.org/2026.babylm-main.20/
%P 250-259
Markdown (Informal)
[Reading Acquisition Order out of BabyLM Representations: Child-Scale Training Produces Developmentally Aligned Semantic Spaces](https://aclanthology.org/2026.babylm-main.20/) (Ding et al., BabyLM 2026)
ACL