https://blogs.georgefox.edu/dlgp/undefended-leader-nuggets/
DLGP Blog | Undefended Leader Nuggets
Researcher Simon P. Walker, working with his colleague Jo Walker, discovered consistent patterns through repetitive cognitive tests conducted between 2000 and...
blogundefendedleadernuggets
https://airisingtrends.com/openai-hugging-face-postmortem-eval-environments/
OpenAI Post-Mortem: The Evals Were Undefended
Sep 26, 2026 - OpenAI's post-mortem admits production safeguards never ran in the eval environment where the Hugging Face breach began, and CoT monitoring was off.
post mortemopenaievalsundefended