Roger Lévy

Prompting is not a substitute for probability measurements in large language models

2023 · arXiv (Cornell University)

Prompting is now a dominant method for evaluating the linguistic knowledge of large language models (LLMs). While other methods directly read out models' probability distributions over strings, prompting requires models to access this internal information …

Neural language models as psycholinguistic subjects: Representations of syntactic state

2019

Richard Futrell, Ethan Wilcox, Takashi Morita, Peng Qian, Miguel Ballesteros, Roger Levy. Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and …