arXiv Machine Learning July 20, 2026 By Jiading Gai, Shuai Zhang, Xiang Song, Bernie Wang, George Karypis

DualKV: Shared-Prompt Flash Attention for Efficient RL Training with Large Rollouts and Long Contexts

Read the original on arXiv Machine Learning →

arXiv:2605. 15422v3 Announce Type: replace Abstract: Modern RL post-training methods such as GRPO and DAPO train on N response sequences of R tokens sampled from a shared prompt of P tokens, but standard FlashAttention replicates all P prompt tokens N times across both forward and backward passes -- duplicating compute and memory on identical hidden states.

Summary generated by The Flow from the publisher's feed. The full article lives at arXiv Machine Learning.

Hugging Face Trending Papers

Jul 23

Windowed-MTP: Removing the Full-Context Draft-KV Tax at Million-Token Context

Speculative decoding accelerates autoregressive generation by having a cheap draft propose tokens that a target verifies in parallel. Frontier models increasingly ship a built-in Multi-Token-Prediction (MTP/NEXTN) draft head under the assumption that the draft is negligibly cheap.

efficiency

arXiv Machine Learning

Jul 24

Windowed-MTP: Removing the Full-Context Draft-KV Tax at Million-Token Context

arXiv:2607. 21535v1 Announce Type: new Abstract: Speculative decoding accelerates autoregressive generation by having a cheap draft propose tokens that a target verifies in parallel.

By Alagappan Valliappan

efficiency

arXiv Machine Learning

Jul 17

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

arXiv:2607. 14952v1 Announce Type: new 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 tokens or below and rely on length generalization at deployment.

By 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

agents reinforcement-learning

arXiv Machine Learning

2d ago

Rollplex: Cross-Phase GPU Spatial Sharing for Vision Language Model Post-Training

arXiv:2608. 14498v1 Announce Type: new Abstract: Vision-language models (VLMs) enable embodied agents to reason and act from visual observations and language instructions.

By Hanfeng Lu, Tianyu Feng, Suyi Li, Yuheng Zhao, Wei Gao, Shaopan Xiong, Ju Huang, Siran Yang, Jiamang Wang, Lin Qu, Wei Wang

llms agents reinforcement-learning robotics multimodal

arXiv AI

Jul 22

GQLA: Group-Query Latent Attention for Hardware-Adaptive Large Language Model Decoding

arXiv:2605. 15250v3 Announce Type: replace-cross Abstract: Multi-head Latent Attention (MLA), the attention used in DeepSeek-V2/V3, jointly compresses keys and values into a low-rank latent and matches the H100 roofline almost perfectly.

By Fanxu Meng

llms efficiency

arXiv AI

Jun 29

Bifocal Diffusion Language Models: Asymmetric Bidirectional Context for Parallel Generation

arXiv:2606. 27732v1 Announce Type: cross Abstract: Discrete diffusion language models (dLLMs) recover masked tokens in parallel, offering significant speedups over autoregressive (AR) generation.

By Yuhang Chen, Xianfeng Wu, Jinhao Duan, Mingfu Liang, Xiaohan Wei, Yunchen Pu, Fei Tian, Chonglin Sun, Parish Aggarwal, Frank Shyu, Luke Simon, Sandeep Pandey, Xi Liu, Tianlong Chen

llms diffusion efficiency benchmarks

Related stories

Windowed-MTP: Removing the Full-Context Draft-KV Tax at Million-Token Context

Windowed-MTP: Removing the Full-Context Draft-KV Tax at Million-Token Context

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

Rollplex: Cross-Phase GPU Spatial Sharing for Vision Language Model Post-Training

GQLA: Group-Query Latent Attention for Hardware-Adaptive Large Language Model Decoding

Bifocal Diffusion Language Models: Asymmetric Bidirectional Context for Parallel Generation