Scalable LLM Math Reasoning Acceleration with Low-rank Distillation

Open in new window