Researchers propose that naturalistic reading time can be predicted using simple n-gram statistics rather than complex language model statistics. They demonstrate this by showing a correlation between n-gram probabilities and eye-tracking-based metrics of reading time on naturalistic text. --- Why it matters: This study matters to researchers in AI because it challenges the assumption that state-of-the-art language models are better at predicting human reading behavior than simple statistical models. Source: https://arxiv.org/abs/2603.09872