How do Transformer Embeddings Represent Compositions? A Functional Analysis
This work addresses the problem of understanding compositionality in AI models for researchers in natural language processing, but it is incremental as it builds on existing methods for analysis.
The study investigated how transformer-based language models represent compound words and found that ridge regression best accounts for compositionality, with vector addition performing nearly as well, and most embedding models are highly compositional while BERT shows poorer performance.
Compositionality is a key aspect of human intelligence, essential for reasoning and generalization. While transformer-based models have become the de facto standard for many language modeling tasks, little is known about how they represent compound words, and whether these representations are compositional. In this study, we test compositionality in Mistral, OpenAI Large, and Google embedding models, and compare them with BERT. First, we evaluate compositionality in the representations by examining six diverse models of compositionality (addition, multiplication, dilation, regression, etc.). We find that ridge regression, albeit linear, best accounts for compositionality. Surprisingly, we find that the classic vector addition model performs almost as well as any other model. Next, we verify that most embedding models are highly compositional, while BERT shows much poorer compositionality. We verify and visualize our findings with a synthetic dataset consisting of fully transparent adjective-noun compositions. Overall, we present a thorough investigation of compositionality.