The owned and the loaned — comparing top novels by holdings vs. ILL (OCLC Research)
This article explores how the list of widely held novels (print holdings) compares with inter‑library loan (ILL) requests, showing how library holdings reflect historical “canon” more than dynamic reading demands
Sonnet or Not, Bot? Poetry Evaluation for Large Models and Datasets
This study tests how well large language models (LLMs) understand poetic form. Features like rhyme, meter, and repetition. Using a dataset of over 4,000 expert-labeled poems, researchers found that LLMs can accurately identify structured forms such as sonnets and sestinas but struggle with free-form or visually based poetry.