@misc{indiciae538c6db490d5,
  title = {Delayed Attention Training Improves Length Generalization in Transformer--RNN Hybrids},
  author = {Buu Phan and Reza Ebrahimi and Sanjay Haresh and Roland Memisevic},
  year = {2025},
  url = {https://arxiv.org/abs/2510.00258},
  note = {Source identifier: 2510.00258}
}