Emerging2d agoNew Research Method Improves AI Model Transparency and Safety Through Self-Consistency TrainingTechnologyScience66%