JED351

Cantonese-adapted language models show stronger predictive fit for human reading

A new study published on arXiv investigates whether language models specifically trained on Cantonese can better predict human reading patterns compared to models trained on Standard Chinese or general-purpose models. R…