How Much Backtracking is Enough? Exploring the Interplay of SFT and RL in Enhancing LLM Reasoning Paper • 2505.24273 • Published 13 days ago • 4 • 3