CoolFace
Modelpublic

mdomina/cybersec-hunt-adversary-emulation-grpo

sourceHugging Faceupdated 4mo agoView on Hugging Face
0likes4downloads
2 commits on main
47cc6164mo ago

GRPO from SFT — 37.5% (same as SFT). Shaped atomic reward did not move it.

mdomina
d8a2e354mo ago

initial commit

mdomina