AIApril 11, 2025Balancing Accuracy and Efficiency in Language Models: A Two-Phase RL Post-Training Approach for Concise Reasoning
