On the Semantics of LM Latent Space: A Vocabulary-defined Approach

Understanding the latent space of language models (LM) is crucial to refining their performance and interpretability. Existing analyses often fall short in providing disentangled (model-centric) insights into LM semantics, and neglect essential aspects of LM adaptation. In response, we introduce a pioneering method called vocabulary-defined semantics, which establishes a reference frame within the LM latent space, ensuring disentangled semantic analysis grounded in LM vocabulary. Our approach transcends prior entangled analysis, leveraging LM vocabulary for model-centric insights. Furthermore, we propose a novel technique to compute logits, emphasising differentiability and local isotropy, and introduce a neural clustering module for semantically calibrating data representations during LM adaptation. Through extensive experiments across diverse text understanding datasets, our approach outperforms state-of-the-art methods of retrieval-augmented generation and parameter-efficient finetuning, showcasing its efficacy and broad applicability. Our findings not only shed light on LM mechanics, but also offer practical solutions to enhance LM performance and interpretability.