Probing Across Time: What Does RoBERTa Know and When?

Zeyu Liu, Yizhong Wang, Jungo Kasai, Hannaneh Hajishirzi, Noah A. Smith


Abstract
Models of language trained on very large corpora have been demonstrated useful for natural language processing. As fixed artifacts, they have become the object of intense study, with many researchers “probing” the extent to which they acquire and readily demonstrate linguistic abstractions, factual and commonsense knowledge, and reasoning abilities. Recent work applied several probes to intermediate training stages to observe the developmental process of a large-scale model (Chiang et al., 2020). Following this effort, we systematically answer a question: for various types of knowledge a language model learns, when during (pre)training are they acquired? Using RoBERTa as a case study, we find: linguistic knowledge is acquired fast, stably, and robustly across domains. Facts and commonsense are slower and more domain-sensitive. Reasoning abilities are, in general, not stably acquired. As new datasets, pretraining protocols, and probes emerge, we believe that probing-across-time analyses can help researchers understand the complex, intermingled learning that these models undergo and guide us toward more efficient approaches that accomplish necessary learning faster.
Anthology ID:
2021.findings-emnlp.71
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2021
Month:
November
Year:
2021
Address:
Punta Cana, Dominican Republic
Editors:
Marie-Francine Moens, Xuanjing Huang, Lucia Specia, Scott Wen-tau Yih
Venue:
Findings
SIG:
SIGDAT
Publisher:
Association for Computational Linguistics
Note:
Pages:
820–842
Language:
URL:
https://aclanthology.org/2021.findings-emnlp.71
DOI:
10.18653/v1/2021.findings-emnlp.71
Bibkey:
Cite (ACL):
Zeyu Liu, Yizhong Wang, Jungo Kasai, Hannaneh Hajishirzi, and Noah A. Smith. 2021. Probing Across Time: What Does RoBERTa Know and When?. In Findings of the Association for Computational Linguistics: EMNLP 2021, pages 820–842, Punta Cana, Dominican Republic. Association for Computational Linguistics.
Cite (Informal):
Probing Across Time: What Does RoBERTa Know and When? (Liu et al., Findings 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.findings-emnlp.71.pdf
Video:
 https://aclanthology.org/2021.findings-emnlp.71.mp4
Code
 leo-liuzy/probe-across-time
Data
BLiMPCoLALAMAMRPCMultiNLIReCoRDSSTSST-2WSC