Evaluating the Zero-shot Robustness of Instruction-tuned Language Models

Published in Submitted to ICLR 2024, 2023