6 papers
I-SDPO: Instance-Level Adaptive Self-Distillation Policy Optimization
Yubo Zhang, Xinhong Ma, Zezhong Tan +1
Group Relative Policy Optimization (GRPO) learns from reward differences within a rollout group, but receives no useful relative signal when every sampled response is incorrect. Pr…
Revisiting Reinforcement Learning with Verifiable Rewards from a Contrastive Perspective
Feng Zhang, Xinhong Ma, Ziqiang Dong +5
Group Relative Policy Optimization (GRPO) is one of the most widely adopted RLVR algorithms for post-training large language models on reasoning tasks. We first show that GRPO admi…
RAVE: Re-Allocating Visual Attention in Large Multimodal Models
Xi Leng, Xinhong Ma, Ziqiang Dong +4
Large multimodal models (LMMs) inherit the self-attention mechanism of pretrained language backbones, yet standard attention can exhibit suboptimal allocation, including cross-moda…
More Edits, More Stable: Understanding the Lifelong Normalization in Sequential Model Editing
Xin Ma, Wei Chen, Qi Liu +4
Lifelong Model Editing aims to continuously update evolving facts in Large Language Models while preserving unrelated knowledge and general capabilities, yet it remains plagued by…
ADHint: Adaptive Hints with Difficulty Priors for Reinforcement Learning
Feng Zhang, Zezhong Tan, Xinhong Ma +6
To address the limited capability expansion and low sample efficiency of Reinforcement Learning (RL), recent methods have integrated ''hints'' into post-training, which are prefix…
Towards Flash Thinking via Decoupled Advantage Policy Optimization
Zezhong Tan, Hang Gao, Xinhong Ma +2
Recent Large Reasoning Models (LRMs) have achieved remarkable performance in solving complex problems via supervised fine-tuning (SFT) and reinforcement learning (RL). Although exi…