Why does surprisal from larger transformer-based language models provide a poorer fit to human reading times?

Cited by (107)

Items 101–107 of 107

1. Large Language Models as Distributional Baselines for Language Tasks
   Sean Trott, James A. Michaelov, Cameron Jones, Tyler A. Chang, Benjamin K. Bergen · 2026 · Open Mind
2. EEG deconvolution modelling of naturalistic reading reveals surprisal effects modulated during mind wandering
   Rosy V. Southwell, Albert E. Kim, Sidney D'Mello · 2026 · Language Cognition and Neuroscience
3. Spillover, depth and articulation: information-theoretic predictors of spoken-word duration in emotional speech
   Jelena Lazic, Sanja Vujnović · 2026 · SSRN Electronic Journal
4. Are Humans and LLMs Confused by the Same Code? An Empirical Study on Fixation-Related Potentials and LLM Perplexity
   Youssef Abdelsalam, Norman Peitek, Anna-Maria Maurer, Mariya Toneva, Sven Apel · 2026 · ACM/IEEE International Conference on Software Engineering: Companion Proceedings (ICSE-Companion)
5. Through the Looking Glass: Eye Tracking as a Cognitive Benchmark for Language-Model Semantics and Observer-Specific Semantic Alignment
   Boris B. Velichkovsky, Ivan G. Mukhin, Indira R. Galeeva, Pavel Yurievich Sereda-Kalinin, Ivan A. Basyul, Yulia V. Bessonova · 2026 · Journal of Eye Movement Research
6. I want to ride my bike/bicycle: predictivity affects word forms in language production but not in language comprehension
   Max S. Dunn, Zhenguang Garry Cai · 2026 · Language and Cognition
7. Semantic Update as a Predictor of Reading Time: Moving Beyond Word-Level Surprisal
   Siddharth Gupta, Alessandro Lopopolo, Milena Rabovsky · 2026 · bioRxiv (Cold Spring Harbor Laboratory)