@misc{indiciae73e7fbcb55c3,
  title = {Adversarially Pretrained Transformers May Be Universally Robust In-Context Learners},
  author = {Soichiro Kumano and Hiroshi Kera and Toshihiko Yamasaki},
  year = {2026},
  url = {https://arxiv.org/abs/2505.14042},
  note = {Source identifier: 2505.14042}
}