Determine whether the Perspective API model has been updated

Determine whether the model underlying the Perspective API has been updated since the experiments reported in the original HateBench study, in order to establish the comparability of current and prior detector evaluations.

Background

The study compares several hate-speech detectors, including the commercial Perspective API, against results from the original HateBench benchmark. Because commercial APIs may change their underlying models without publicly documenting each update, differences between current and historical performance cannot necessarily be attributed to detector improvements or degradation unless the model version is known.

The authors explicitly identify the update status of the Perspective API model as unresolved, which affects the reproducibility and interpretation of comparisons with the original experiments.