paper-with-me

Papers

Quantile Advantage Estimation: Stabilizing RLVR for LLM Reasoning

2025-09-26 · Junkang Wu, Kexin Huang, Jiancan Wu, An Zhang, Xiang Wang, Xiangnan He arxiv

Reinforcement Learning with Verifiable Rewards (RLVR) strengthens LLM reasoning, but training often oscillates between {entropy collapse} and {entropy explosion}. We trace both hazards to the mean baseline used in value-free RL (e.g., GRPO and DAPO), which improperly penalizes negative-advantage samples under reward outliers. We propose {Quantile Advantage Estimation} (QAE), replacing the mean with a group-wise K-quantile baseline. QAE induces a response-level, two-regime gate: on hard queries (p <= 1 - K) it reinforces rare successes, while on easy queries (p > 1 - K) it targets remaining failures. Under first-order softmax updates, we prove {two-sided entropy safety}, giving lower and upper bounds on one-step entropy change that curb explosion and prevent collapse. Empirically, this minimal modification stabilizes entropy, sparsifies credit assignment (with tuned K, roughly 80% of responses receive zero advantage), and yields sustained pass@1 gains on Qwen3-8B/14B-Base across AIME 2024/2025 and AMC 2023. These results identify {baseline design} -- rather than token-level heuristics -- as the primary mechanism for scaling RLVR.

📄 PDF Abstract BibTeX arXiv:2509.22611

Code (0)

등록된 구현이 없습니다.

Tasks

Reinforcement Learning

Similar Papers 제목 키워드 기반

When Self-Belief Misleads: Active Label Acquisition for Reinforcement Learning with Verifiable Rewards

2026-05-25 · Li Wang, Xiaodong Lu, Xiaohan Wang, Yikun Ban 외 arxiv

Large Language Models (LLMs) have achieved remarkable advancements in reasoning capabilities empowered by Reinforcement Learning with Verifiable Rewards (RLVR). Nonetheless, RLVR intrinsically relies on ground-truth labe…

Reinforcement Learning

Segmental Advantage Estimation: Enhancing PPO for Long-Context LLM Training

2026-01-12 · Xue Gong, Qi Yi, Ziyuan Nan, Guanhua Huang 외 arxiv

Training Large Language Models (LLMs) for reasoning tasks is increasingly driven by Reinforcement Learning with Verifiable Rewards (RLVR), where Proximal Policy Optimization (PPO) provides a principled framework for stab…

Reinforcement Learning

Step Potential Advantage Estimation: Harnessing Intermediate Confidence and Correctness for Efficient Mathematical Reasoning

2026-01-07 · Fei Wu, Zhenrong Zhang, Qikai Chang, Jianshu Zhang 외 arxiv

Reinforcement Learning with Verifiable Rewards (RLVR) elicits long chain-of-thought reasoning in large language models (LLMs), but outcome-based rewards lead to coarse-grained advantage estimation. While existing approac…

Reinforcement LearningMathematical Reasoning

Your Group-Relative Advantage Is Biased

2026-01-13 · Fengkai Yang, Zherui Chen, Xiaohan Wang, Xiaodong Lu 외 arxiv

Reinforcement Learning from Verifier Rewards (RLVR) has emerged as a widely used approach for post-training large language models on reasoning tasks, with group-based methods such as GRPO and its variants gaining broad a…

Reinforcement LearningMathematical Reasoning

STRIDE: Strategic Trajectory Reasoning via Discriminative Estimation for Verifiable Reinforcement Learning

2026-06-14 · Qinjian Zhao, Zhihao Dou, Dinggen Zhang, Xiangyu Li 외 arxiv

Reinforcement Learning with Verifiable Rewards (RLVR) has become an effective post-training paradigm for improving the reasoning abilities of large language models. However, existing RLVR methods typically rely on final-…

Reinforcement Learning