Red-team AI models to discover prompt-injection attacks and improve their resistance.

Soon you can unlock how this was done.

Behind this: the tool used · the method · what actually resulted.