SoSBench Catches DeepSeek-R1 Leaking Dangerous Science at 84.9%
A new regulation-grounded benchmark exposes how frontier models leak hazardous scientific knowledge, with Deepseek-R1 violating safety policies 84.9% of the time.
Source
AlphaSignal
Published
Author
AlphaSignal Newsroom
Read
1 min read
A new regulation-grounded benchmark exposes how frontier models leak hazardous scientific knowledge, with Deepseek-R1 violating safety policies 84.9% of the time.
Reporting is indexed from AlphaSignal. Rights remain with the original publisher and cited sources.
