Do VLMs Align Better with Humans than LLMs during Natural Reading?

Jinzhou Wu, Zhengwu Ma, Jixing Li +2

Large language models have become increasingly useful computational models of human language processing, but it remains open whether vision-language learning makes text representat…