@article{sparrowsparserolloutforstableandefficien, title = {Sparrow: Sparse Rollout for Stable and Efficient Long-context RL of Large Language Models}, author = {Yang Zhou and Ranajoy Sadhukhan and Zhaofeng Sun and Zhuoming Chen and Souvik Kundu and Saket Dingliwal and Sai Muralidhar Jayanthi and Aram Galstyan and Haizhong Zheng and Beidi Chen}, year = {2026}, eprint = {2606.08446}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2606.08446}, }