RBS-Attention: Radius-Bounded Sparse Prefill for Long-Context Large Language Models
arXiv:2609.20971v1 Announce Type: new Abstract: Long-context large language model inference is increasingly limited by prefill, where dense self-attention processes the entire prompt before generation begins. Sparse block selection can...
Read more: https://arxiv.org/abs/2609.20971
Scan with your phone
Scan with your phone