Which Word Orders Facilitate Length Generalization in LMs? An Investigation with GCG-Based Artificial Languages

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: El-Naggar, Nadine, Kuribayashi, Tatsuki, Briscoe, Ted
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912647050428416
author El-Naggar, Nadine
Kuribayashi, Tatsuki
Briscoe, Ted
author_facet El-Naggar, Nadine
Kuribayashi, Tatsuki
Briscoe, Ted
contents Whether language models (LMs) have inductive biases that favor typologically frequent grammatical properties over rare, implausible ones has been investigated, typically using artificial languages (ALs) (White and Cotterell, 2021; Kuribayashi et al., 2024). In this paper, we extend these works from two perspectives. First, we extend their context-free AL formalization by adopting Generalized Categorial Grammar (GCG) (Wood, 2014), which allows ALs to cover attested but previously overlooked constructions, such as unbounded dependency and mildly context-sensitive structures. Second, our evaluation focuses more on the generalization ability of LMs to process unseen longer test sentences. Thus, our ALs better capture features of natural languages and our experimental paradigm leads to clearer conclusions -- typologically plausible word orders tend to be easier for LMs to productively generalize.
format Preprint
id arxiv_https___arxiv_org_abs_2510_12722
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Which Word Orders Facilitate Length Generalization in LMs? An Investigation with GCG-Based Artificial Languages
El-Naggar, Nadine
Kuribayashi, Tatsuki
Briscoe, Ted
Computation and Language
Whether language models (LMs) have inductive biases that favor typologically frequent grammatical properties over rare, implausible ones has been investigated, typically using artificial languages (ALs) (White and Cotterell, 2021; Kuribayashi et al., 2024). In this paper, we extend these works from two perspectives. First, we extend their context-free AL formalization by adopting Generalized Categorial Grammar (GCG) (Wood, 2014), which allows ALs to cover attested but previously overlooked constructions, such as unbounded dependency and mildly context-sensitive structures. Second, our evaluation focuses more on the generalization ability of LMs to process unseen longer test sentences. Thus, our ALs better capture features of natural languages and our experimental paradigm leads to clearer conclusions -- typologically plausible word orders tend to be easier for LMs to productively generalize.
title Which Word Orders Facilitate Length Generalization in LMs? An Investigation with GCG-Based Artificial Languages
topic Computation and Language
url https://arxiv.org/abs/2510.12722