Discriminator for evaluating fake text

[zainsarwar865]

Hello,
In your paper there was a section titled Empirical validation in which you compute AUC using multiple tests. As I understand, you trained a classifier using features generated from GLTR tool. Coud you provide more details about the kind of classifier used, how much training was done etc? I am trying to replicate the results in the paper for my research.

Thank you!

[izhx]

Same questions here.

[zainsarwar865]

Hey izhx, they may not provide this but we reproduced their models in our paper. You can find the code and paper here : https://github.com/jmpu/DeepfakeTextDetection