Hacker News new | ask | show | jobs
CAD: Disaggregating Core Attention for Efficient Long-Context LLM Training (hao-ai-lab.github.io)
6 points by ginda307 184 days ago