Tag
This paper introduces a checkpoint-wise activation-intervention framework to study how concepts become functionally sufficient during language model training, comparing preservation targets across layers and models.
This paper situates large language models within the broader history of computational approaches to concept analysis in the history, philosophy, and sociology of science (HPSS), reviewing methodological challenges and LLM-based case studies for lexical semantic change detection. It covers corpus construction, operationalization, and evaluation across both pre-LLM and LLM-era workflows.