AI RESEARCH
Offline Exploration-Aware Fine-Tuning for Long-Chain Mathematical Reasoning
arXiv CS.LG
•
ArXi:2603.16206v1 Announce Type: new Through encouraging self-exploration, reinforcement learning from verifiable rewards (RLVR) has significantly advanced the mathematical reasoning capabilities of large language models. As the starting point for RLVR, the capacity of supervised fine-tuning (SFT) to memorize new chain-of-thought trajectories provides a crucial initialization that shapes the subsequent exploration landscape. However, existing research primarily focuses on facilitating exploration during RLVR