English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Past studies have provided broad support for that words with lowerpredictability (i.e., higher surprisal) require more time for comprehension byusing large language models (LLMs) to simulate humans' cognitive load. Ingeneral, these studies have implicitly assumed that the probability scores fromLLMs are accurate, ignoring the discrepancies between human cognition and LLMsfrom this standpoint. Inspired by the concept of probability calibration, weare the first work to focus on the probability distribution for human readingsimulation. We propose to use temperature-scaled surprisal, a surprisalcalculated by shaped probability, to be the predictor of human reading times.Our results across three corpora consistently revealed that such a surprisalcan drastically improve the prediction of reading times. Setting thetemperature to be approximately 2.5 across all models and datasets can yield upto an 89% of increase in delta log-likelihood in our setting. We also propose acalibration metric to quantify the possible human-likeness bias. Furtheranalysis was done and provided insights into this phenomenon.

Improving fit to human reading times via temperature-scaled surprisal