arXiv:2604. 20140v2 Announce Type: replace Abstract: Direct Preference Optimization (DPO) is an effective framework for aligning large language models with human preferences, but it struggles with complex reasoning tasks.
By Darsh Kachroo, Arjun Prasaath Anbazhagan, Adriana Caraeni, Brennan Lagasse, Kevin Zhu
arXiv:2609.15664v1 Announce Type: cross
Abstract: Amid the rapid advancement of physical-world intelligence, cloud-edge collaborative large language models (LLMs) have emerged as a promising roadmap...
By Victor H. Chen, Hairui Yu, Stella K. Chung, Hong Yan
arXiv:2606. 01561v1 Announce Type: new Abstract: Aligning Large Language Models (LLMs) with human preferences is often formulated via Direct Preference Optimization (DPO).
By Xiwen Chen, Wenhui Zhu, Jingjing Wang, Peijie Qiu, Zhipeng Wang, Huayu Li, ZhengXiao He, Xuanzhao Dong, Prayag Tiwari, Mingkun Xu, Yujian Xiong, Feng Luo, Abolfazl Razi, Brendan Hogan Rappazzo, Anderson Schneider, Yuriy Nevmyvaka
arXiv:2509. 23982v2 Announce Type: replace-cross Abstract: Preference alignment is a critical step in making Large Language Models (LLMs) useful and aligned with (human) preferences.
By Lucio La Cava, Andrea Tagarelli
The paper argues that in offline preference optimization for reasoning models, applying gradients uniformly to all chosen–rejected pairs is inefficient and can be harmful. It introduces the concept of gradient utility, showing that a pair’s contribution depends on both informativeness and stability, and that high-gradient samples often lie in high‑curvature regions, causing noisy updates. To address this, the authors propose SAGE (Stability‑Aware Gradient Efficiency), which maintains difficulty‑stratified candidate pools and selects only high‑utility pairs for backpropagation, resulting in smoother optimization and better performance on mathematical reasoning benchmarks.
By Hui Wu, Hengyi Cai, Jinman Zhao, Xinran Chen, Ziheng Li, Zhejun Zhao, Shuaiqiang Wang, Yuchen Li, Dawei Yin
arXiv:2606. 04503v1 Announce Type: cross Abstract: Reinforcement learning with verifiable rewards (RLVR) has greatly advanced large reasoning models (LRMs), but it requires timely training on a huge fully-annotated dataset.
By Guangcheng Zhu, Shenzhi Yang, Haobo Wang, Xing Zheng, Yingfan MA, Xuening Feng, Zhongqi Chen, Bowen Song, Weiqiang Wang, Gang Chen