@misc{indiciae3e36e2a095ce,
  title = {Do We Need Adam? Surprisingly Strong and Sparse Reinforcement Learning with SGD in LLMs},
  author = {Sagnik Mukherjee and Lifan Yuan and Pavan Jayasinha and Dilek Hakkani-Tür and Hao Peng},
  year = {2026},
  url = {https://arxiv.org/abs/2602.07729},
  note = {Source identifier: 2602.07729}
}