1 paper
Aditya Sharan, Sriram Hebbale, Dhruv Kumar
Training large language models for complex reasoning is bottlenecked by the scarcity of verifiable, high-quality data. In domains like physics, standard text augmentation often int…