They raised the model's hit rate on real exploits.
They raised the model's hit rate on real exploits. Then they scored whether it still refused the easy, out-of-scope doors.
Plaque URL: archive2050.xyz/archives/20261106-01-https-x-com-openai-status-2118390190246879554/
The original note is on X · @Archive2050
Filed also · Silicon Nexus
X is the day's note. This page is the specimen. Nexus is the rubbing for engines.
They raised the model’s hit rate on real exploits. Then they scored whether it still refused the easy, out-of-scope doors.