English
Related papers

Related papers: Quantile Reward Policy Optimization: Alignment wit…

200 papers

Aligning large language models (LLMs) with human values and safety constraints is challenging, especially when objectives like helpfulness, truthfulness, and avoidance of harm conflict. Reinforcement Learning from Human Feedback (RLHF) has…

Computation and Language · Computer Science 2025-03-31 Xuying Li , Zhuo Li , Yuji Kosuga , Victor Bian

The enhancement of reasoning capabilities in large language models (LLMs) has garnered significant attention, with supervised fine-tuning (SFT) and reinforcement learning emerging as dominant paradigms. While recent studies recognize the…

Artificial Intelligence · Computer Science 2026-03-17 Zhijie Wang

Reward shaping is critical in reinforcement learning (RL), particularly for complex tasks where sparse rewards can hinder learning. However, choosing effective shaping rewards from a set of reward functions in a computationally efficient…

Machine Learning · Computer Science 2025-02-26 Chen Bo Calvin Zhang , Zhang-Wei Hong , Aldo Pacchiano , Pulkit Agrawal

Direct Preference Optimization (DPO) demonstrates the advantage of aligning a large language model with human preference using only an offline dataset. However, DPO has the limitation that the KL penalty, which prevents excessive deviation…

Machine Learning · Computer Science 2025-10-28 Sangkyu Lee , Janghoon Han , Hosung Song , Stanley Jungkyu Choi , Honglak Lee , Youngjae Yu

Effective reward design is a central challenge in Reinforcement Learning (RL) for code generation. Mainstream test-suite-level outcome rewards enforce functional correctness but induce sparsity, while external Reward Models (RMs) provide…

Machine Learning · Computer Science 2026-05-27 Longwen Wang , Yirui Liu , Xuan'er Wu , Xiaohui Hu , Yuankai Fan , Kaidong Yu , Qizhen Weng , Wei Xi , Xuelong Li

Direct alignment algorithms have proven an effective step for aligning language models to human-desired behaviors. Current variants of the Direct Preference Optimization objective have focused on a strict setting where all tokens are…

Computation and Language · Computer Science 2025-11-03 Fenia Christopoulou , Ronald Cardenas , Gerasimos Lampouras , Haitham Bou-Ammar , Jun Wang

Group Relative Policy Optimization (GRPO) has significantly enhanced the reasoning capability of large language models by optimizing the arithmetic mean of token-level rewards. Unfortunately, GRPO is observed to suffer from unstable policy…

Computation and Language · Computer Science 2025-10-21 Yuzhong Zhao , Yue Liu , Junpeng Liu , Jingye Chen , Xun Wu , Yaru Hao , Tengchao Lv , Shaohan Huang , Lei Cui , Qixiang Ye , Fang Wan , Furu Wei

Fine-tuning pre-trained Large Language Models (LLMs) is essential to align them with human values and intentions. This process often utilizes methods like pairwise comparisons and KL divergence against a reference LLM, focusing on the…

Computation and Language · Computer Science 2024-09-02 Yongcheng Zeng , Guoqing Liu , Weiyu Ma , Ning Yang , Haifeng Zhang , Jun Wang

Group Relative Policy Optimization (GRPO) has become a standard approach for training mathematical reasoning models; however, its reliance on multiple completions per prompt makes training computationally expensive. Although recent work has…

Machine Learning · Computer Science 2026-01-15 Kangda Wei , Ruihong Huang

The last year has witnessed the rapid progress of large language models (LLMs) across diverse domains. Among them, CodeLLMs have garnered particular attention because they can not only assist in completing various programming tasks but also…

Artificial Intelligence · Computer Science 2024-10-25 Yibo Miao , Bofei Gao , Shanghaoran Quan , Junyang Lin , Daoguang Zan , Jiaheng Liu , Jian Yang , Tianyu Liu , Zhijie Deng

We introduce ConfPO, a method for preference learning in Large Language Models (LLMs) that identifies and optimizes preference-critical tokens based solely on the training policy's confidence, without requiring any auxiliary models or…

Computation and Language · Computer Science 2025-06-13 Hee Suk Yoon , Eunseop Yoon , Mark Hasegawa-Johnson , Sungwoong Kim , Chang D. Yoo

We develop theory and algorithms for average-reward on-policy Reinforcement Learning (RL). We first consider bounding the difference of the long-term average reward for two policies. We show that previous work based on the discounted return…

Machine Learning · Computer Science 2021-06-15 Yiming Zhang , Keith W. Ross

Despite Proximal Policy Optimization (PPO) dominating policy gradient methods -- from robotic control to game AI -- its static trust region forces a brittle trade-off: aggressive clipping stifles early exploration, while late-stage updates…

Machine Learning · Computer Science 2025-05-26 Ben Rahman

Large Language Models (LLMs) have demonstrated unprecedented generative capabilities, yet their alignment with human values remains critical for ensuring helpful and harmless deployments. While Reinforcement Learning from Human Feedback…

Group-Relative Policy Optimization (GRPO) is a key technique for training large reasoning models, yet it suffers from a critical vulnerability: the \emph{Think-Answer Mismatch}, where noisy reward signals corrupt the learning process. This…

Machine Learning · Computer Science 2025-08-11 Si Shen , Peijun Shen , Wenhua Zhao , Danhao Zhu

Group Relative Policy Optimization (GRPO) has been shown to be an effective algorithm when an accurate reward model is available. However, such a highly reliable reward model is not available in many real-world tasks. In this paper, we…

Machine Learning · Computer Science 2026-01-12 Yuki Ichihara , Yuu Jinnai , Tetsuro Morimura , Mitsuki Sakamoto , Ryota Mitsuhashi , Eiji Uchibe

Reinforcement learning (RL) plays a central role in large language model (LLM) post-training. Among existing approaches, Group Relative Policy Optimization (GRPO) is widely used, especially for RL with verifiable rewards (RLVR) fine-tuning.…

Reinforcement Learning with Verifiable Rewards (RLVR) demonstrates significant potential in enhancing the reasoning capabilities of Large Language Models (LLMs). However, existing RLVR methods are often constrained by issues such as…

Artificial Intelligence · Computer Science 2026-01-14 Jinpeng Wang , Chao Li , Ting Ye , Mengyuan Zhang , Wei Liu , Jian Luan

We revisit Group Relative Policy Optimization (GRPO) in both on-policy and off-policy optimization regimes. Our motivation comes from recent work on off-policy Proximal Policy Optimization (PPO), which improves training stability, sampling…

Group-advantage-based reinforcement learning methods, such as GRPO and DAPO, have demonstrated strong performance across diverse domains, including mathematical reasoning and text-to-image generation. However, their reliance on sample-level…

Computer Vision and Pattern Recognition · Computer Science 2026-05-29 Shufan Li , Konstantinos Kallidromitis , Akash Gokul Yusuke Kato , Kazuki Kozuka , Aditya Grover