@misc{indiciaec456a35e31ca,
  title = {Sample-Efficient Online Learning in LM Agents via Hindsight Trajectory Rewriting},
  author = {Michael Y. Hu and Benjamin Van Durme and Jacob Andreas and Harsh Jhamtani},
  year = {2026},
  url = {https://arxiv.org/abs/2510.10304},
  note = {Source identifier: 2510.10304}
}