Recent advances in large language model (LLM) reasoning have shown that sophisticated behaviors such as planning and self-reflection can emerge through reinforcement learning (RL). However, despite these successes, RL in its current form remains insufficient to induce capabilities that exceed the limitations of the base model, as it is primarily optimized based on existing knowledge of the model rather than facilitating the acquisition of new information. To address this limitation, we employ supervised fine-tuning (SFT) to learn what RL cannot, which enables the incorporation of new knowledge and reasoning patterns by leveraging high-quality demonstration data. We analyze the training dynamics of RL and SFT for LLM reasoning and find that RL excels at maintaining and improving performance on questions within the model's original capabilities, while SFT is more effective at enabling progress on questions beyond the current scope of the model. Motivated by the complementary strengths of RL and SFT, we introduce a novel training approach, \textbf{ReLIFT} (\textbf{Re}inforcement \textbf{L}earning \textbf{I}nterleaved with Online \textbf{F}ine-\textbf{T}uning). In ReLIFT, the model is primarily trained using RL, but when it encounters challenging questions, high-quality solutions are collected for fine-tuning, and the training process alternates between RL and fine-tuning to enhance the model's reasoning abilities. ReLIFT achieves an average improvement of over +5.2 points across five competition-level benchmarks and one out-of-distribution benchmark compared to other zero-RL models. Furthermore, we demonstrate that ReLIFT outperforms both RL and SFT while using only 13\% of the detailed demonstration data, highlighting its scalability. These results provide compelling evidence that ReLIFT overcomes the fundamental limitations of RL and underscores the significant potential.

翻译：大型语言模型（LLM）推理领域的最新进展表明，诸如规划和自我反思等复杂行为可以通过强化学习（RL）涌现。然而，尽管取得了这些成功，当前形式的强化学习仍不足以诱导出超越基础模型局限性的能力，因为它主要基于模型现有知识进行优化，而非促进新信息的获取。为应对这一局限，我们采用监督微调（SFT）来学习强化学习无法掌握的内容，从而通过利用高质量演示数据来整合新知识和推理模式。我们分析了强化学习和监督微调在大型语言模型推理中的训练动态，发现强化学习擅长维持和改进模型原有能力范围内问题的表现，而监督微调在推动模型解决超出当前能力范围的问题方面更为有效。受强化学习与监督微调互补优势的启发，我们提出了一种新颖的训练方法——\textbf{ReLIFT}（\textbf{Re}inforcement \textbf{L}earning \textbf{I}nterleaved with Online \textbf{F}ine-\textbf{T}uning）。在ReLIFT中，模型主要使用强化学习进行训练，但当遇到具有挑战性的问题时，会收集高质量解决方案进行微调，训练过程在强化学习与微调之间交替进行，以增强模型的推理能力。与其他零强化学习模型相比，ReLIFT在五个竞赛级基准测试和一个分布外基准测试上平均提升了超过+5.2分。此外，我们证明ReLIFT在仅使用13%的详细演示数据的情况下，性能优于纯强化学习和监督微调，突显了其可扩展性。这些结果提供了有力证据，表明ReLIFT克服了强化学习的根本局限性，并彰显了其巨大潜力。