Derivational Probing: Unveiling the Layer-wise Derivation of Syntactic Structures in Neural Language Models
This work addresses the need to interpret internal representations in language models for researchers in NLP and linguistics, but it is incremental as it builds on existing probing methods.
The paper tackled the problem of understanding how syntactic structures are constructed across layers in neural language models, revealing a bottom-up derivation where micro-syntactic structures emerge in lower layers and integrate into macro-syntactic structures in higher layers, with experiments on BERT showing that the timing of this integration is critical for downstream performance.
Recent work has demonstrated that neural language models encode syntactic structures in their internal representations, yet the derivations by which these structures are constructed across layers remain poorly understood. In this paper, we propose Derivational Probing to investigate how micro-syntactic structures (e.g., subject noun phrases) and macro-syntactic structures (e.g., the relationship between the root verbs and their direct dependents) are constructed as word embeddings propagate upward across layers. Our experiments on BERT reveal a clear bottom-up derivation: micro-syntactic structures emerge in lower layers and are gradually integrated into a coherent macro-syntactic structure in higher layers. Furthermore, a targeted evaluation on subject-verb number agreement shows that the timing of constructing macro-syntactic structures is critical for downstream performance, suggesting an optimal timing for integrating global syntactic information.