@misc{indiciaeb824ebf1da07,
  title = {Understanding LLM Performance Degradation in Multi-Instance Processing: The Roles of Instance Count and Context Length},
  author = {Jingxuan Chen and Mohammad Taher Pilehvar and Jose Camacho-Collados},
  year = {2026},
  url = {https://arxiv.org/abs/2603.22608},
  note = {Source identifier: 2603.22608}
}