CE-GPPO: Coordinating Entropy via Gradient-Preserving Clipping Policy Optimization in Reinforcement Learning (2026.acl-long)
Copied to clipboard
| Challenge: | Existing methods for proximal policy optimization discard valuable gradient signals from low-probability tokens due to the clipping mechanism. |
| Approach: | They propose an algorithm that reintroduces gradients from clipped tokens in native PPO in a gentle and bounded manner. |
| Outcome: | The proposed algorithm outperforms strong baselines on reasoning benchmarks on different model scales. |
Similar Papers
Entropy Ratio Clipping as a Soft Global Constraint for Stable Reinforcement Learning (2026.findings-acl)
Copied to clipboard
Zhenpeng Su, Leiyu Pan, Minxuan Lv, Tiehua Mei, Zijia Lin, Yuntao Li, Wenping Hu, Ruiming Tang, Kun Gai, Guorui Zhou
| Challenge: | Large language model post-training often adopts an off-policy training paradigm . however, the off-poliicy training model introduces distribution shifts that push the policy beyond the trust region. |
| Approach: | They propose to use the entropy ratio as a global metric to measure the relative change in policy exploration throughout updates. |
| Outcome: | Experiments show that the proposed metric improves performance across multiple benchmarks. |
Heterogeneous Adaptive Policy Optimization: Tailoring Optimization to Every Token’s Nature (2026.acl-long)
Copied to clipboard
| Challenge: | Existing methods that use entropy as a discrete filter or post-hoc regulator are limited in their ability to optimize for reasoning tasks. |
| Approach: | They propose a token-aware algorithm that continuously adapts optimization dynamics based on token-level entropy throughout the entire training process. |
| Outcome: | Extensive experiments on mathematical reasoning, code, and logic tasks across multiple models demonstrate HAPO’s consistent superiority over DAPO. |
From log 𝜋 to 𝜋: Taming Divergence in Soft Clipping via Bilateral Decoupled Decay of Probability Gradient Weight (2026.acl-long)
Copied to clipboard
Xiaoliang Fu, Jiaye Lin, Yangyi Fang, Chaowen Hu, Cong Qin, Zekai Shao, Binbin Zheng, Lu Pan, Ke Zeng
| Challenge: | Standard algorithms for Large Language Models (LLMs) enforce stability via "hard clipping" but relying on log-probability gradient yields divergent weights as probabilities vanish, destabilizing LLM training. |
| Approach: | They propose a decoupled gradient policy optimization that uses a decay mechanism to decouple the probability of a boundary token. |
| Outcome: | The proposed algorithm outperforms baselines on various mathematical benchmarks. |
Targeted Exploration via Unified Entropy Control for Reinforcement Learning (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing methods for group relative policy optimization suffer from entropy collapse . Existing exploration methods introduce additional bias or variance during exploration, making it difficult to maintain stability. |
| Approach: | They propose a framework that provides targeted mechanisms for exploration and stabilization. |
| Outcome: | The proposed framework expands search space on difficult prompts while preventing entropy growth uncontrollably. |
Adapt to Thrive! Adaptive Power-Mean Policy Optimization for Improved LLM Reasoning (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing methods for reinforcement learning with verifiable rewards (RLVR) rely on static objective functions and rigid clipping strategies that misalign with the model’s evolving reasoning capabilities. |
| Approach: | They propose to incorporate Power-Mean Policy Optimization (PMPO) and Feedback-Adaptive Clipping (FAC) to overcome limitations of static mechanisms. |
| Outcome: | Extensive experiments on nine reasoning tasks show the proposed paradigm outperforms state-of-the-art methods. |
MASPO: Unifying Gradient Utilization, Probability Mass, and Signal Reliability for Robust and Sample-Efficient LLM Reasoning (2026.acl-long)
Copied to clipboard
Xiaoliang Fu, Jiaye Lin, Yangyi Fang, Binbin Zheng, Chaowen Hu, Zekai Shao, Cong Qin, Lu Pan, Ke Zeng, Xunliang Cai
| Challenge: | Existing RLVR algorithms rely on rigid, uniform, and symmetric trust region mechanisms . current algorithms lack robustness, asymmetric signal reliability and inefficient gradient utilization . |
| Approach: | They propose a framework to harmonize three dimensions of RLVR algorithms, a paper argues . a binary cutoff is used to discard valuable reinforcement signals, they argue . |
| Outcome: | The proposed framework outperforms baselines in evaluating a robust RLVR solution. |
Taming Extreme Tokens: Covariance-Aware GRPO with Gaussian-Kernel Advantage Reweighting (2026.acl-short)
Copied to clipboard
| Challenge: | Excessive exploitation can cause the model to become overconfident in its suboptimal solutions, thereby limiting its capabilities to explore novel reasoning strategies. |
| Approach: | They propose a method that dynamically down-weights extreme token-level updates via a Gaussian kernel and reduces the instability caused by the trade-off. |
| Outcome: | The proposed method improves downstream performance across reasoning benchmarks and stabilizes entropy as training progresses. |
LEPO: Latent Reasoning Policy Optimization for Large Language Models (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing latent reasoning methods that use chain of thought (CoT) are limited to selecting one discrete token at each reasoning step, which potentially induces information loss. |
| Approach: | They propose a framework that injects controllable stochasticity into latent reasoning via Gumbel-Softmax, restoring LLMs' exploratory capacity and enhancing their compatibility with Reinforcement Learning (RL). |
| Outcome: | The proposed framework preserves richer information for more comprehensive reasoning and is compatible with Reinforcement Learning (RL). |
Beyond High-Entropy Exploration: Correctness-Aware Low-Entropy Segment-Based Advantage Shaping for Reasoning LLMs (2026.findings-acl)
Copied to clipboard
| Challenge: | Recent work studies RLVR through token entropy, arguing that high-entropies drive exploration and should receive stronger updates. |
| Approach: | They propose a correctness-aware reinforcement framework that performs fine-grained advantage modulation over low-entropy segments. |
| Outcome: | The proposed framework improves accuracy over strong RL baselines across three backbones and six math benchmarks while maintaining high-entropy exploration. |
Understanding and Preventing Entropy Collapse in RLVR with On-Policy Entropy Flow Optimization (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing RLVR algorithms suffer from entropy collapse, leading to premature determinism and unstable optimization. |
| Approach: | They propose an adaptive entropy flow balancing mechanism that rescales entropic-increasing and enotro-decreazing updates according to their contributions to enthroy change. |
| Outcome: | The proposed method outperforms existing RLVR algorithms on six reasoning benchmarks. |