Model Internal Sleuthing: Finding Lexical Identity and Inflectional Features in Modern Language Models

arXiv:2506.02132v5 Announce Type: replace Abstract: Large transformer-based language models dominate modern NLP, yet our understanding of how they encode linguistic information relies primarily on studies of early models like BERT and GPT-2. We systematically probe 25 models from BERT Base to Qwen2.5-7B focusing on two linguistic properties: lexical identity and inflectional features across 6 diverse languages. We find a consistent pattern: inflectional features are linearly decodable throughout the model, while lexical identity is prominent early but increasingly weakens with depth. Further analysis of the representation geometry reveals that models with aggressive mid-layer dimensionality compression show reduced steering effectiveness in those layers, despite probe accuracy remaining high. Pretraining analysis shows that inflectional structure stabilizes early while lexical identity representations continue evolving. Taken together, our findings suggest that transformers maintain inflectional features across layers, while trading off lexical identity for compact, predictive representations. Our code is available at https://github.com/ml5885/model_internal_sleuthing

Related

• Sparse or Dense? A Mechanistic Estimation of Computation Density in Transformer-based LLMs
• Cell-Based Representation of Relational Binding in Language Models
• Do LLMs Encode Functional Importance of Reasoning Tokens?
• Vocab Diet: Reshaping the Vocabulary of LLMs via Vector Arithmetic

Source: arXiv cs.CL | 2026-04-23