让大模型自己指导探索,基于梯度方向优化推理能力。
Can LLMs Guide Their Own Exploration? Gradient-Guided Reinforcement Learning for LLM Reasoning
- 用模型自身梯度几何引导探索,而非外部启发式方法。
- 在多个数学和通用推理任务上显著提升通过率与主答案率。
- 探索方向更正交且保持语义连贯,适合强化学习优化场景。
强化学习已成为提升大语言模型推理能力的关键手段,但现有探索机制与模型实际学习过程严重脱节。熵奖励和外部语义比较器仅促进表面变化,无法保证采样轨迹在影响优化的方向上存在差异。我们提出G2RL框架,其探索由模型自身的梯度更新几何驱动。对于每个响应,G2RL从模型最后一层敏感性构建序列级特征,仅需标准前向传播即可获取,并通过比较采样组内特征来衡量各轨迹对策略的重塑效果。引入新梯度方向的轨迹获得有界乘法奖励,冗余或脱离流形的更新则被弱化,从而生成自指型探索信号,天然契合PPO风格的稳定性和KL控制。在Qwen3 base 1.7B和4B模型上,针对MATH500、AMC、AIME24、AIME25、GPQA、MMLUpro等数学与通用推理基准测试,G2RL在pass@1、maj@16、pass@k指标上持续优于基于熵的GRPO和外部嵌入方法。分析诱导出的几何结构发现,G2RL将探索扩展至更正交甚至相反的梯度方向,同时维持语义连贯性,表明模型自身的更新空间是指导大模型强化学习探索更忠实有效的基础。
原文摘要 · Abstract (English)
Reinforcement learning has become essential for strengthening the reasoning abilities of large language models, yet current exploration mechanisms remain fundamentally misaligned with how these models actually learn. Entropy bonuses and external semantic comparators encourage surface level variation but offer no guarantee that sampled trajectories differ in the update directions that shape optimization. We propose G2RL, a gradient guided reinforcement learning framework in which exploration is driven not by external heuristics but by the model own first order update geometry. For each response, G2RL constructs a sequence level feature from the model final layer sensitivity, obtainable at negligible cost from a standard forward pass, and measures how each trajectory would reshape the policy by comparing these features within a sampled group. Trajectories that introduce novel gradient directions receive a bounded multiplicative reward scaler, while redundant or off manifold updates are deemphasized, yielding a self referential exploration signal that is naturally aligned with PPO style stability and KL control. Across math and general reasoning benchmarks (MATH500, AMC, AIME24, AIME25, GPQA, MMLUpro) on Qwen3 base 1.7B and 4B models, G2RL consistently improves pass@1, maj@16, and pass@k over entropy based GRPO and external embedding methods. Analyzing the induced geometry, we find that G2RL expands exploration into substantially more orthogonal and often opposing gradient directions while maintaining semantic coherence, revealing that a policy own update space provides a far more faithful and effective basis for guiding exploration in large language model reinforcement learning.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品;中文卡片由大模型生成,请以原文为准。