TY  - RPRT
TI  - Why Does Surprisal From Larger Transformer-Based Language Models Provide a Poorer Fit to Human Reading Times?
AU  - Byung-Doh Oh
AU  - William Schuler
PY  - 2022
UR  - https://arxiv.org/abs/2212.12131
ID  - 2212.12131
ER  -