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.
Original language | English |
---|---|
Title of host publication | Findings of the Association for Computational Linguistics: EMNLP 2023 |
Place of Publication | Singapore |
Publisher | Association for Computational Linguistics |
Publication date | 6 Dec 2023 |
Pages | 13190-13208 |
Publication status | Published - 6 Dec 2023 |
Keywords
- representational spaces
- language modeling
- NLP
- information theoretic probing
- cross-task similarity
- linguistic information emergence
- syntactic knowledge
- semantic learning
- reasoning tasks
- model interpretability