Two papers (FedCLIP and ChatGPT robustness) are accepted by ICLR 2023 workshop on robust large models! [Papers]