ResearchSimon Willison

Quoting Anthropic Frontier Red Team

#ai#security#research#glm#anthropic

English

The Anthropic Frontier Red Team evaluated several AI models on a binary exploitation benchmark, finding that GLM-5.3 achieved control flow hijacks in 4% of trials, while Claude Mythos Preview succeeded in 6%. This indicates a significant advancement in AI capabilities compared to earlier models, which failed to achieve any hijacks.

中文

Anthropic Frontier Red Team对多个AI模型进行了二进制利用基准测试，发现GLM-5.3在4%的试验中实现了控制流劫持，而Claude Mythos Preview则在6%的试验中成功。这表明与早期模型相比，AI能力有了显著提升，后者未能实现任何劫持。