We investigate whether large language models encode latent knowledge of frame semantics, focusing on frame identification, a core challenge in frame semantic parsing that involves selecting the appropriate semantic frame for a target word in context. Using the FrameNet lexical resource, we evaluate models under prompt-based inference and observe that they can perform frame identification effectively even without explicit supervision. To assess the impact of task-specific training, we fine-tune the model on FrameNet data, which substantially improves in-domain accuracy while generalizing well to out-of-domain benchmarks. Further analysis shows that the models can generate semantically coherent frame definitions, highlighting the model's internalized understanding of frame semantics.
@article{arxiv.2509.19540,
title = {Do LLMs Encode Frame Semantics? Evidence from Frame Identification},
author = {Jayanth Krishna Chundru and Rudrashis Poddar and Jie Cao and Tianyu Jiang},
journal= {arXiv preprint arXiv:2509.19540},
year = {2026}
}