Repository with experiments for the paper: "A statistical measure of trust in generative AI".
Experiments for three domains are in the corresponding folders. To reproduce them, please consequently run all files in each folder step-by-step.
The interactive results mentioned in the paper are presented in folder INTERACTIVE_RESULTS. All results have the following names: <domain>_<classificator>_<optional:labelling/generator>_dashboard.html. It is recommended to download the files and open in browser.