A Better Way to Do Masked Language Model Scoring

Carina Kauf, Anna Ivanova


Abstract
Estimating the log-likelihood of a given sentence under an autoregressive language model is straightforward: one can simply apply the chain rule and sum the log-likelihood values for each successive token. However, for masked language models (MLMs), there is no direct way to estimate the log-likelihood of a sentence. To address this issue, Salazar et al. (2020) propose to estimate sentence pseudo-log-likelihood (PLL) scores, computed by successively masking each sentence token, retrieving its score using the rest of the sentence as context, and summing the resulting values. Here, we demonstrate that the original PLL method yields inflated scores for out-of-vocabulary words and propose an adapted metric, in which we mask not only the target token, but also all within-word tokens to the right of the target. We show that our adapted metric (PLL-word-l2r) outperforms both the original PLL metric and a PLL metric in which all within-word tokens are masked. In particular, it better satisfies theoretical desiderata and better correlates with scores from autoregressive models. Finally, we show that the choice of metric affects even tightly controlled, minimal pair evaluation benchmarks (such as BLiMP), underscoring the importance of selecting an appropriate scoring metric for evaluating MLM properties.
Anthology ID:
2023.acl-short.80
Volume:
Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)
Month:
July
Year:
2023
Address:
Toronto, Canada
Venue:
ACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
925–935
Language:
URL:
https://aclanthology.org/2023.acl-short.80
DOI:
10.18653/v1/2023.acl-short.80
Bibkey:
Cite (ACL):
Carina Kauf and Anna Ivanova. 2023. A Better Way to Do Masked Language Model Scoring. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 925–935, Toronto, Canada. Association for Computational Linguistics.
Cite (Informal):
A Better Way to Do Masked Language Model Scoring (Kauf & Ivanova, ACL 2023)
Copy Citation:
PDF:
https://preview.aclanthology.org/remove-xml-comments/2023.acl-short.80.pdf