11 papers · 1 filter
Co-RL: Unsupervised Reasoning Emerges from Diverse Cohort in Multi-agent RL
Yunhao Yang, Yuexin Bian, Yunjie Tian +6
Reinforcement learning (RL) has emerged as a powerful approach for improving reasoning in language and vision-language models, yet its strongest successes still depend heavily on g…
Learn What's Left, Not What's Mastered: Saturation Aware Advantage Reweighting for Multi-Reward Policy Optimization
Yixuan Wang, Yifei Chen, Haichao Zhang +6
Reinforcement learning (RL) with group-relative advantages has become the de facto standard for post-training language model reasoners. However, when optimizing multiple reward obj…
Self-Supervised Visual On-Policy Distillation
Yijiang Li, Yijun Liang, Yunjie Tian +6
Visual on-policy distillation relies heavily on an informative teacher-student asymmetry, through either a larger, stronger teacher or privileged supervision, such as reference ans…
Demystifying Agent Skills: Why They Work-Until They Don't
Zhiyuan Jiang, Fangrui Huang, Hanwen Xing +6
Skills have emerged as a practical and effective approach for enhancing LLM agents at inference time through structured packages of knowledge. However, existing evaluations largely…
On-Policy Self-Distillation without Any Supervision
Yijiang Li, Bingyang Wang, Yijun Liang +3
On-policy (Self-)Distillation (OPD / OPSD) has shown strong potential for post-training large language models (LLMs). However, existing methods still rely heavily on external super…
Visual Contrastive Self-Distillation
Yijun Liang, Yunjie Tian, Yijiang Li +5
On-policy self-distillation (OPSD) removes the external teacher required by on-policy distillation (OPD), but still needs asymmetric information between teacher and student so that…
From RGB Generation to Dense Field Readout: Pixel-Space Dense Prediction with Text-to-Image Models
Zanyi Wang, Xin Lin, Haodong Li +2
Large-scale text-to-image models are attractive backbones for dense prediction because RGB generation pretraining learns rich semantic, structural, and geometric priors. Existing g…
Anisotropic Modality Align
Xiaomin Yu, Yijiang Li, Yuhui Zhang +8
Training multimodal large language models has long been limited by the scarcity of high-quality paired multimodal data. Recent studies show that the shared representation space of…
SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Tasks
Xiangyi Li, Yimin Liu, Wenbo Chen +75
Agent Skills are structured packages of procedural knowledge that augment large language model (LLM) agents at inference time. Despite rapid adoption, there is no standard way to m…
Less is Enough: Synthesizing Diverse Data in LLM Feature Space with Sparse Autoencoders
Zhongzhi Li, Xuansheng Wu, Yijiang Li +2
The diversity of post-training data is critical for effective downstream performance in large language models (LLMs). Many existing approaches to constructing post-training data qu…
RN-D: Discretized Categorical Actors for On-Policy Reinforcement Learning
Yuexin Bian, Jie Feng, Tao Wang +3
On-policy Reinforcement Learning (RL) remains a dominant paradigm for continuous control, yet standard implementations rely on Gaussian actors and relatively shallow MLP policies,…