Scalable LLM Math Reasoning Acceleration with Low-rank Distillation

LRM
Re-assign community