Researcher — paper.dinq.me

2026 顶会论文

作者位次 #8/8ICML 2026Poster· Deep Learning / Large Language Models

MoFO: Momentum-Filtered Optimizer for Mitigating Forgetting in LLM Fine-Tuning

Yupeng Chen · Senmiao Wang · Yushun Zhang · Zhihang Lin · Haozhe Zhang · Weijian Sun · Tian Ding · Ruoyu Sun

强化学习与后训练OpenReview

作者位次 #5/8ICML 2026Poster· Deep Learning / Large Language Models

Knapsack RL: Compute-Efficient Reinforcement Learning via Heterogeneous Rollout Allocation

Ziniu Li · Congliang Chen · Tianyun Yang · Tian Ding · Ruoyu Sun · Ge Zhang · Wenhao Huang · Zhiquan Luo

强化学习与后训练效率与系统OpenReview

作者位次 #4/5ACL 2026Poster· Natural Language Processing

VCORE : Variance-Controlled Optimization-based Reweighting for Chain-of-Thought Supervision

Xuan Gong · Senmiao Wang · Hanbo Huang · Ruoyu Sun · Shiyu Liang

摘要

Supervised fine-tuning (SFT) on long chain-of-thought (CoT) trajectories has emerged as a crucial technique for enhancing the reasoning abilities of large language models (LLMs). However, the standard cross-entropy loss treats all tokens equally, ignoring their heterogeneous contributions across...

推理强化学习与后训练评测与基准OpenReview 项目页

Ruoyu Sun

获取完整研究档案

联系方式

2026 顶会论文

MoFO: Momentum-Filtered Optimizer for Mitigating Forgetting in LLM Fine-Tuning

Knapsack RL: Compute-Efficient Reinforcement Learning via Heterogeneous Rollout Allocation

VCORE : Variance-Controlled Optimization-based Reweighting for Chain-of-Thought Supervision

常见合作者