arXiv 2610.06647 · cs.CL · 2026-10-05
LoGRA: Scaling LLM Reinforcement Learning with Low-Rank Gradient Sketches
Shaokun Zhang, Yifan Zhang, Jian Hu, Yueying Li, Hao Zhang, Binfeng Xu, Jan Kautz, Yi Dong
NVIDIA
In latest · #139 2026-W41 #8 Added on 2026-10-21
One-line takeaway
LoGRA cuts RL post-training memory with low-rank gradient sketches and keeps training stable with predicted-KL step control.
LoGRA 用低秩梯度草图压缩 RL 后训练的显存，并按预测 KL 控制步长保持稳定。
Read on arXiv PDF Hugging Face discussion
Topics & disciplines
More on Reinforcement learning
- Waypoint-OPD: Transferring RL-Induced Policy Shifts via Gated On-Policy Distillation ▲218
- Self-Retrospection Distillation: Turning Post-hoc Experiences into Prior Foresight ▲189
- Sharpening Tax in Post-Training ▲111
- RULER: Instance-aware Rubric Rewards for SVG Generation ▲104
- TPPO: Tracking-Value PPO For Language Model Reasoning ▲96
About this data
Metadata from the arXiv API (verified); upvotes from Hugging Face Daily Papers; citations from Semantic Scholar. Last refreshed 2026-10-22.
The takeaway is written by AIMap from the abstract; the abstract itself is not reproduced — read the original.