Reasoning Path Compression: Compressing Generation Trajectories for Efficient LLMReasoning

Open in new window