English
Related papers

Related papers: MASPO: Unifying Gradient Utilization, Probability …

200 papers

Reinforcement Learning with Verifiable Rewards (RLVR) is commonly based on group sampling to estimate advantages and stabilize policy updates. In practice, computational limits often rule out very large groups, so training proceeds with…

Reinforcement Learning with Verifiable Rewards (RLVR) enhances Large Language Model (LLM) reasoning but suffers from advantage collapse on ``hard samples'' where all rollouts fail. This lack of variance eliminates crucial learning signals.…

Machine Learning · Computer Science 2026-05-08 Xinyu Lu , Kaiqi Zhang , Jinglin Yang , Boxi Cao , Yaojie Lu , Hongyu Lin , Min He , Xianpei Han , Le Sun

Reinforcement Learning with Verifiable Rewards (RLVR) has advanced the reasoning capabilities of Large Language Models (LLMs) by leveraging direct outcome verification instead of learned reward models. Building on this paradigm, Group…

Machine Learning · Computer Science 2026-04-23 Jingyi Wang , Lei Zhu , Tengjin Weng , Song-Li Wu , Haochen Tan , Jierun Chen , Chaofan Tao , Haoli Bai , Lu Hou , Lifeng Shang , Xiao-Ping Zhang

Reinforcement Learning with Verifiable Rewards (RLVR) has achieved remarkable success in improving autoregressive models, especially in domains requiring correctness like mathematical reasoning and code generation. However, directly…

Machine Learning · Computer Science 2026-03-03 Chenxing Wei , Jiazhen Kang , Hong Wang , Jianqing Zhang , Hao Jiang , Xiaolong Xu , Ningyuan Sun , Ying He , F. Richard Yu , Yao Shu , Bo Jiang

Group Relative Policy Optimization (GRPO) was introduced and used recently for promoting reasoning in LLMs under verifiable (binary) rewards. We show that the mean + variance calibration of these rewards induces a weighted contrastive loss…

Machine Learning · Computer Science 2025-10-22 Youssef Mroueh

Reinforcement Learning with Human Feedback (RLHF) has been the dominant approach for improving the reasoning capabilities of Large Language Models (LLMs). Recently, Reinforcement Learning with Verifiable Rewards (RLVR) has simplified this…

Computation and Language · Computer Science 2025-10-10 Yining Wang , Jinman Zhao , Chuangxin Zhao , Shuhao Guan , Gerald Penn , Shinan Liu

Reinforcement Learning with Verifiable Rewards (RLVR) has demonstrated remarkable success in enhancing the reasoning capabilities of Large Language Models (LLMs). However, recent studies question whether RL genuinely expands reasoning…

Machine Learning · Computer Science 2026-02-04 Dayu Wang , Jiaye Yang , Weikang Li , Jiahui Liang , Yang Li

Reinforcement Learning with Verifiable Rewards (RLVR) has proven to be a highly effective strategy for endowing Large Language Models (LLMs) with robust multi-step reasoning abilities. However, its design and optimizations remain tailored…

Computation and Language · Computer Science 2026-04-15 Zhenhailong Wang , Xuehang Guo , Sofia Stoica , Haiyang Xu , Hongru Wang , Hyeonjeong Ha , Xiusi Chen , Yangyi Chen , Ming Yan , Fei Huang , Heng Ji

Reinforcement Learning with Verifiable Rewards (RLVR) improves LLM reasoning, yet growing evidence indicates an exploration ceiling: it often reweights existing solution traces rather than discovering new strategies, limiting gains under…

Machine Learning · Computer Science 2026-03-03 Bizhe Bai , Xinyue Wang , Peng Ye , Tao Chen

RLVR has enhanced the reasoning capabilities of Large Language Models (LLMs) across various tasks. However, GRPO, a representative RLVR algorithm, suffers from a critical limitation: when all responses within a group are either entirely…

Machine Learning · Computer Science 2025-09-24 Gongrui Nan , Siye Chen , Jing Huang , Mengyu Lu , Dexun Wang , Chunmei Xie , Weiqi Xiong , Xianzhou Zeng , Qixuan Zhou , Yadong Li , Xingzhong Xu

A significant portion of recent research on Large Language Model (LLM) alignment focuses on developing new policy optimization methods based on Group Relative Policy Optimization (GRPO). Two prominent directions have emerged: (i) a shift…

Machine Learning · Computer Science 2026-02-27 Svetlana Glazyrina , Maksim Kryzhanovskiy , Roman Ischenko

Reinforcement learning with verifiable rewards (RLVR) has recently enhanced the reasoning capabilities of large language models (LLMs), particularly for mathematical problem solving. However, a fundamental limitation remains: as the…

Machine Learning · Computer Science 2025-11-03 Wenhao Deng , Long Wei , Chenglei Yu , Tailin Wu

Multimodal large language models (MLLMs) have demonstrated impressive reasoning and instruction-following capabilities, yet their expanded modality space introduces new compositional safety risks that emerge from complex text-image…

Cryptography and Security · Computer Science 2025-11-18 Xuankun Rong , Wenke Huang , Tingfeng Wang , Daiguo Zhou , Bo Du , Mang Ye

Large Language Models (LLMs) are increasingly deployed in business-critical domains such as finance, education, healthcare, and customer support, where users expect consistent and reliable recommendations. Yet LLMs often exhibit variability…

Machine Learning · Computer Science 2026-04-20 Sonal Prabhune , Balaji Padmanabhan , Kaushik Dutta

Group Relative Policy Optimization (GRPO) is highly effective for post-training autoregressive (AR) language models, yet its direct application to diffusion large language models (dLLMs) often triggers reward collapse. We identify two…

Machine Learning · Computer Science 2026-03-10 Jianyuan Zhong , Kaibo Wang , Ding Ding , Zijin Feng , Haoli Bai , Yang Xiang , Jiacheng Sun , Qiang Xu

Differentiable reinforcement learning (RL) frameworks like DiffRO offer a powerful approach for controllable text-to-speech (TTS), but are vulnerable to reward hacking, particularly for nuanced tasks like emotion control. The policy model…

Sound · Computer Science 2026-02-17 Cong Wang , Changfeng Gao , Yang Xiang , Zhihao Du , Keyu An , Han Zhao , Qian Chen , Xiangang Li , Yingming Gao , Ya Li

Reinforcement Learning from Verifiable Rewards (RLVR) significantly enhances large language models (LLMs) reasoning but severely suffers from calibration degeneration, where models become excessively over-confident in incorrect answers.…

Machine Learning · Computer Science 2026-05-28 Zhengzhao Ma , Xueru Wen , Boxi Cao , Yaojie Lu , Hongyu Lin , Jinglin Yang , Min He , Xianpei Han , Le Sun

Post-training LLMs with Reinforcement Learning, specifically Group Relative Policy Optimization (GRPO), has emerged as a paradigm for enhancing mathematical reasoning. However, standard GRPO relies on scalar correctness rewards that are…

Computation and Language · Computer Science 2026-03-03 Xiwen Chen , Wenhui Zhu , Peijie Qiu , Xuanzhao Dong , Hao Wang , Haiyu Wu , Huayu Li , Aristeidis Sotiras , Yalin Wang , Abolfazl Razi

Optimizing discrete diffusion model (DDM) with rewards remains a challenge: the non-autoregressive paradigm makes importance sampling intractable and rollout complex, puzzling reinforcement learning methods such as Group Relative Policy…

Artificial Intelligence · Computer Science 2025-10-06 Tianren Ma , Mu Zhang , Yibing Wang , Qixiang Ye

Current critic-less RLHF methods aggregate multi-objective rewards via an arithmetic mean, leaving them vulnerable to constraint neglect: high-magnitude success in one objective can numerically offset critical failures in others (e.g.,…

Machine Learning · Computer Science 2026-05-08 Ivan Montero , Tomasz Jurczyk , Bhuwan Dhingra