Reality Defender’s text detector detects model generated text by observing many pairs of human-written text and model-generated text. It reads the entire input text as a whole and extracts statistical patterns by projecting the text into a high-dimensional vector space. This allows our model to pick up on subtle signals that are hard for humans to detect, and enables it to make accurate and robust predictions when detecting model-generated texts.
How does Reality Defender detect generative text?
Written by Diana Hsieh
Updated over a year ago