Subspace Chronicles: How Linguistic Information Emerges, Shifts and Interacts during Language Model Training

Publikation: Konference artikel i Proceeding eller bog/rapport kapitelKonferencebidrag i proceedingsForskningpeer review

Abstract

Representational spaces learned via language modeling are fundamental to Natural Language Processing (NLP), however there has been limited understanding regarding how and when during training various types of linguistic information emerge and interact. Leveraging a novel information theoretic probing suite, which enables direct comparisons of not just task performance, but their representational subspaces, we analyze nine tasks covering syntax, semantics and reasoning, across 2M pre-training steps and five seeds. We identify critical learning phases across tasks and time, during which subspaces emerge, share information, and later disentangle to specialize. Across these phases, syntactic knowledge is acquired rapidly after 0.5% of full training. Continued performance improvements primarily stem from the acquisition of open-domain knowledge, while semantics and reasoning tasks benefit from later boosts to long-range contextualization and higher specialization. Measuring cross-task similarity further reveals that linguistically related tasks share information throughout training, and do so more during the critical phase of learning than before or after. Our findings have implications for model interpretability, multi-task learning, and learning from limited data.
OriginalsprogEngelsk
TitelFindings of the Association for Computational Linguistics: EMNLP 2023
UdgivelsesstedSingapore
ForlagAssociation for Computational Linguistics
Publikationsdato6 dec. 2023
Sider13190-13208
StatusUdgivet - 6 dec. 2023

Fingeraftryk

Dyk ned i forskningsemnerne om 'Subspace Chronicles: How Linguistic Information Emerges, Shifts and Interacts during Language Model Training'. Sammen danner de et unikt fingeraftryk.

Citationsformater