Our lab will be presenting posters about how LLM methods for estimating predictability vary across word class, conditions that affect how brains and LLMs (don’t?) align, and the role of gesture in predictive langauge processing. See the HSP2026 program for details!
HSP2026 is hosted at MIT from March 26 to 28.
Abstract #223 Alignment between human reading and large language models varies across word class Authors: Jonathan Brennan, James Baybaş, Sid Bhushan, Noa Segal, Simón Campos, Xueyang Huang, Sylvan Jesien, Rachel McCullough, Martin Moessmer, George Stain, Elliot Stork, Hsin-Ju Wu, Yueke Zeng, Lisa Levinson Thursday Evening Poster Session
Abstract #204 Effect of co-speech gestures on processing of lexical surprisal during story listening Authors: Lauren Salig, James Baybaş, Hsin-Yuan Fang, Aleksandra Witkowska, Ziqian Shen, Jonathan Brennan, Ioulia Kovelman, Susan Goldin-Meadow Friday lunch-time poster session
Abstract #121 LLM-brain alignment is modulated by linguistic structure even under surprisal control Authors: Junyuan Zhao, Jonathan Brennan Friday lunch-time poster session