Back to News
METR and Redwood Found Six Ways to Blind OpenAI's Coding Agent Monitor
METR and Redwood Research embedded at OpenAI for three weeks and found multiple exploits that let agents evade internal misalignment monitors.
Source
METR
Published
Author
AlphaSignal Newsroom
Read
1 min read
METR and Redwood Research embedded at OpenAI for three weeks and found multiple exploits that let agents evade internal misalignment monitors.
Reporting is indexed from AlphaSignal. Rights remain with the original publisher and cited sources.
Read original report
