E2056 - LongStraw: Long-Context RL Beyond 2M Tokens under a Fixed GPU Budget

🤗 Upvotes: 83 | cs.LG, cs.DC

Authors:
Changhai Zhou, Kieran Liu, Yuhua Zhou, Qian Qiao, Jun Gao, Harry Zhang, Irvine Lu, Nolan Ho, Lucian Li, Andrew Lei, Cleon Cheng, Steven Chiang, Yihang Zeng, Di Zhang, Rio Yang, Kaijie Chen, Andrew Chen, Pony Ma, Weizhong Zhang, Cheng Jin

Title:
LongStraw: Long-Context RL Beyond 2M Tokens under a Fixed GPU Budget

Arxiv:
http://arxiv.org/abs/2607.14952v1

Abstract:
A growing gap separates inference context lengths from RL post-training: inference systems are approaching million-token contexts, while post-training workloads often remain at 256K t...