TY  - RPRT
TI  - To model human linguistic prediction, make LLMs less superhuman
AU  - Byung-Doh Oh
AU  - Tal Linzen
PY  - 2026
UR  - https://arxiv.org/abs/2510.05141
ID  - 2510.05141
ER  -