Long-context RLVR is powerful, but most of the cost comes from generating massive CoT deployments. Sparse attention would make it faster, that's it.
Sparrow: Sparse Deployment for Long-Context RL LLM
By
–
By
–
Long-context RLVR is powerful, but most of the cost comes from generating massive CoT deployments. Sparse attention would make it faster, that's it.