Is word order considered by foundation models? A comparative task-oriented analysis
Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt
Standard
Is word order considered by foundation models? A comparative task-oriented analysis. / Zhao, Qinghua; Li, Jiaang; Liu, Junfeng; Kang, Zhongfeng; Zhou, Zenghui.
I: Expert Systems with Applications, Bind 241, 122700, 2024.Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - JOUR
T1 - Is word order considered by foundation models? A comparative task-oriented analysis
AU - Zhao, Qinghua
AU - Li, Jiaang
AU - Liu, Junfeng
AU - Kang, Zhongfeng
AU - Zhou, Zenghui
N1 - Publisher Copyright: © 2023 Elsevier Ltd
PY - 2024
Y1 - 2024
N2 - Word order, a linguistic concept essential for conveying accurate meaning, is seemingly not that necessary in language models based on the existing works. Contrary to this prevailing notion, our paper delves into the impacts of word order by employing carefully selected tasks that demand distinct abilities. Using three large language model families (ChatGPT, Claude, LLaMA), three controllable word order perturbation strategies, one novel perturbation qualification metric, four well-chosen tasks, and three languages, we conduct experiments to shed light on this topic. Empirical findings demonstrate that Foundation models take word order into consideration during generation. Moreover, tasks emphasizing reasoning abilities exhibit a greater reliance on word order compared to those primarily based on world knowledge.
AB - Word order, a linguistic concept essential for conveying accurate meaning, is seemingly not that necessary in language models based on the existing works. Contrary to this prevailing notion, our paper delves into the impacts of word order by employing carefully selected tasks that demand distinct abilities. Using three large language model families (ChatGPT, Claude, LLaMA), three controllable word order perturbation strategies, one novel perturbation qualification metric, four well-chosen tasks, and three languages, we conduct experiments to shed light on this topic. Empirical findings demonstrate that Foundation models take word order into consideration during generation. Moreover, tasks emphasizing reasoning abilities exhibit a greater reliance on word order compared to those primarily based on world knowledge.
KW - Foundation model
KW - MGSM
KW - Order perturbation ratio
KW - WinoGrande
KW - Word order
U2 - 10.1016/j.eswa.2023.122700
DO - 10.1016/j.eswa.2023.122700
M3 - Journal article
AN - SCOPUS:85178663660
VL - 241
JO - Expert Systems with Applications
JF - Expert Systems with Applications
SN - 0957-4174
M1 - 122700
ER -
ID: 378947148