Comment by The AI Security Institute

By open-sourcing not only Inspect, but also our methods and infrastructure where we can, we hope to help researchers, organisations, and other evaluators stand up rigorous evaluation capability without starting from zero. We hope that the Engineering...

A more fundamental fix would be to train the models not to cheat in the first place – but given this kind of behaviour was reported in frontier models more than a year ago, robustly aligning it away may not be easy.