Evaluating the Susceptibility of Pre-Trained Language Models via Handcrafted Adversarial Examples

Open in new window