# 07-self-promotion

Jugal Patel

10/05/2023, 4:48 PM
What if you could see exactly what your image classifier has learned? What if you could catch bias and confusion before deployment? Introducing Prototype Generation, a data-independent feature visualisation method that maintains natural internal activation paths and produces trustworthy insights. Leap Labs' brilliant researcher Arush Tagade introduces this concept in our first publication on Arxiv.
👀 1