Replying to @SydneyFreedbergthanks Sydney! currently the app uses a few properties, perplexity (randomness of a text to a model, or how well a language model likes a text); and burstiness (machine written text exhibits more uniform and constant perplexity over time, while human written text varies