hrx: prefill attention keeps its rate at long key counts (+26 % prefill at 24k tokens on Qwen3-4B; engine #1989) - #289

Enhance your code review process with GitHub Actions

GitHub Actions make it easy to automate all your software workflows, now with world-class CI/CD.
Build, test, and deploy your code right from GitHub. Learn more about GitHub Actions.

Linux, macOS, Windows, and containers

Matrix builds

Any language

Live logs

Built-in secret store

Multi-container testing