AdaptEvo is a framework for learning under imperfect supervision in rule-governed contextual decision tasks. It couples confidence‑adaptive policy optimization (CA‑GRPO) with an evolution module that synthesizes reusable decision knowledge from recurring failures and refines process rubrics. On an industrial multimodal content moderation dataset, AdaptEvo outperforms baseline GRPO and fixed reward mixtures, achieving higher exact‑label and binary decision accuracy both in‑period and out‑of‑period.
By Shijun Wan, Jiancong Xie, Hang Xu, Jin Duan, Qixiong Wang, Xi Xiang, Maofei Que, Yahui Liu, Zhongyu Wei, Mu Chuan
arXiv:2610.02824v1 Announce Type: new
Abstract: Rubric-based reinforcement learning extends reward-driven optimization to open-ended tasks by assigning partial credit to individual response requireme...
By Yuxuan Fan, Jaehong Yoon
arXiv:2606. 08480v1 Announce Type: cross Abstract: Reinforcement learning (RL) presents a promising avenue for enhancing generative recommendation beyond supervised imitation, leveraging reward signals to guide policy improvement.
By Kewei Xu, Junbo Qi, Yanyan Zou, Pengfei Zhang, Xingzhi Yao, Shengjie Li
arXiv:2609.06107v1 Announce Type: new
Abstract: Data policies for reinforcement learning with verifiable rewards (RLVR) determine which rollouts are used, how strongly they are weighted, and which do...
By Hao Liang, Mingrui Chen, Hengyi Feng, Meiyi Qiang, Wentao Zhang
arXiv:2608.30005v1 Announce Type: new
Abstract: Rubric-based reinforcement learning extends RL beyond tasks with exact answers or rule-based verifiers by scoring responses against instance-specific c...
By Fengyu Xie, Yilun Zhao, Bingsen Chen, Arman Cohan, Chen Zhao
TextualRL is a gradient‑free reinforcement learning framework that updates the textual context guiding large language model agents without altering their parameters. It treats groups of rollouts from the same task as units of critique, contrasting successful and failed runs to generate context patches, and validates these patches using held‑out rewards. Experiments on six benchmarks show that TextualRL boosts test accuracy by 24.4 and 15.3 percentage points for GPT‑5.5 and Qwen3.8‑27B, respectively.
By Tianyu Huai, Jie Zhou, Bo Mao, Qin Chen, Liang He, Xipeng Qiu
arXiv:2606. 03980v1 Announce Type: new Abstract: Reward models (RMs) provide critical feedback signals for LLM post-training, notably in reinforced fine-tuning (RFT) and reinforcement learning (RL) pipelines.
By Tao Chen, Gangwei Jiang, Pengyu Cheng, Siyuan Huang, Yihao Liu, Jingwei Ni, Jiaqi Guo, Mengyu Zhou, Kai Tang, Junling Liu, Qinliang Su, Xiaoxi Jiang, Guanjun Jiang
arXiv:2606. 05784v1 Announce Type: new Abstract: We identify and formally characterize credit misassignment as a systematic failure mode of GRPO in tool-augmented multimodal search agents: its uniform broadcast of trajectory-level advantages to all tokens causes valuable tool-use steps in failing trajectories to be penalized no differently from valueless ones.
By Chengqi Dong, Chuhuai Yue, Hang He, yandong liu, Fenghe Tang, S Kevin Zhou, Xiaohan Wang, Jiajun Chai, Guojun Yin
arXiv:2607. 25659v1 Announce Type: new Abstract: Rubric-based reinforcement learning enriches language model training by evaluating model outputs against explicit criteria.
By Bo-Wen Zhang, Junwei He, Wen Wang, Song-Lin Lv, Wentao Ma, Rongyi Lin, Shuhan Zhong, Lan-Zhe Guo
Large language model (LLM) agents require post-training methods that can improve long-horizon decision making from environment feedback. However, existing agentic post-training pipelines often treat data curation as a fixed preprocessing step, focusing mainly on data augmentation while neglecting filtering, refinement, and adaptation to downstream failures.
LLM agents act in external environments where each action changes the state that later decisions condition on, and where a single wrong step can waste interaction budget or trigger irreversible side effects long before the final failure is observed. Reliable deployment therefore requires \emph{step-level confidence estimation}: a calibrated probability that each proposed action is productive, available \emph{before} the action is executed.
The paper introduces Cross-Behavior Policy Optimization (CBPO), a method that unifies offline demonstrations and policy-generated rollouts into a single regression loss using a fixed-reference target. CBPO eliminates the need for separate supervised fine‑tuning and reinforcement learning objectives, allowing positive and negative offline completions, reference responses, and current and historical rollouts to share a common reward‑labeled format and replay buffer. Experiments on two model scales and four mathematical reasoning benchmarks show that CBPO, especially the CBPO‑Unified variant, outperforms baseline methods in avg@32 and pass@32 metrics, with historical rollouts providing additional gains.
By Du Cheng, Junyu Lu, Jingfan Zhang, Yuzhuo Zhang, Haoxuan Jia, Shichao Weng