We investigate the semantic knowledge of language models (LMs), focusing on (1) whether these LMs create categories of linguistic environments based on their semantic monotonicity properties, and (2) whether these categories play a similar role in LMs as in human language understanding, using negative polarity item licensing as a case study. We introduce a series of experiments consisting of probing with diagnostic classifiers (DCs), linguistic acceptability tasks, as well as a novel DC ranking method that tightly connects the probing results to the inner workings of the LM. By applying our experimental pipeline to LMs trained on various filtered corpora, we are able to gain stronger insights into the semantic generalizations that are acquired by these models.1

Language Models Use Monotonicity to Assess NPI Licensing / Jumelet, J.; Denic, M.; Szymanik, J.; Hupkes, D.; Steinert-Threlkeld, S.. - (2021), pp. 4958-4969. (Intervento presentato al convegno ACL tenutosi a Bangkok nel 2021) [10.18653/v1/2021.findings-acl.439].

Language Models Use Monotonicity to Assess NPI Licensing

Szymanik, J.;
2021-01-01

Abstract

We investigate the semantic knowledge of language models (LMs), focusing on (1) whether these LMs create categories of linguistic environments based on their semantic monotonicity properties, and (2) whether these categories play a similar role in LMs as in human language understanding, using negative polarity item licensing as a case study. We introduce a series of experiments consisting of probing with diagnostic classifiers (DCs), linguistic acceptability tasks, as well as a novel DC ranking method that tightly connects the probing results to the inner workings of the LM. By applying our experimental pipeline to LMs trained on various filtered corpora, we are able to gain stronger insights into the semantic generalizations that are acquired by these models.1
2021
Findings of the Association of Computational Linguistics
USA
Association of Computational Linguistics
Jumelet, J.; Denic, M.; Szymanik, J.; Hupkes, D.; Steinert-Threlkeld, S.
Language Models Use Monotonicity to Assess NPI Licensing / Jumelet, J.; Denic, M.; Szymanik, J.; Hupkes, D.; Steinert-Threlkeld, S.. - (2021), pp. 4958-4969. (Intervento presentato al convegno ACL tenutosi a Bangkok nel 2021) [10.18653/v1/2021.findings-acl.439].
File in questo prodotto:
File Dimensione Formato  
2105.13818.pdf

accesso aperto

Tipologia: Post-print referato (Refereed author’s manuscript)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 1.24 MB
Formato Adobe PDF
1.24 MB Adobe PDF Visualizza/Apri
2021.findings-acl.439.pdf

accesso aperto

Tipologia: Versione editoriale (Publisher’s layout)
Licenza: Creative commons
Dimensione 1.23 MB
Formato Adobe PDF
1.23 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11572/371621
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 11
  • ???jsp.display-item.citation.isi??? ND
social impact