The paper introduces BALIGN, a balanced data selection strategy designed to reduce catastrophic forgetting—referred to as the alignment tax—in large language models during preference-based alignment. By analyzing preference optimization gradients, the authors identify three data-centric features that influence parameter drift: the reference model's log-probability margin, token length differences between chosen and rejected responses, and TF‑IDF similarity to general capability corpora. BALIGN aggregates these features into a composite risk score to filter out high-risk preference samples, thereby preserving foundational capabilities while maintaining alignment gains with minimal computational overhead.
By Minsu Kim, Jianxun Lian, Xing Xie, Steven Euijong Whang
arXiv:2504. 06659v2 Announce Type: replace-cross Abstract: Despite advances in Preference Alignment (PA) for Large Language Models (LLMs), mainstream methods like reinforcement learning with human feedback face notable challenges.
By Xiaohua Feng, Yuyuan Li, Huwei Ji, Jiaming Zhang, Li Zhang, Tianyu Du, Chaochao Chen
arXiv:2608.23149v1 Announce Type: cross
Abstract: The alignment of Large Language Models heavily relies on English-centric high-quality preference data, which often leads to suboptimal performance in...
By Seungyoon Lee, Minhyuk Kim, Jungseob Lee, Heuiseok Lim
arXiv:2607. 03248v1 Announce Type: cross Abstract: The alignment of large language models with human preferences is commonly achieved through Reinforcement Learning from Human Feedback or Direct Preference Optimization.
By Jialiang Wang, Xianming Liu, Xiong Zhou, Hui Liu, Haoliang Li
The paper investigates how preference tuning—optimizing language models with explicit preference signals—behaves when applied to new domains. It systematically compares five alignment objectives and several adaptation strategies, such as target‑domain supervised fine‑tuning and pseudo‑labeling, across summarization, question‑answering helpfulness, and safety tasks. Results show that while pseudo‑labeling reduces domain‑shift degradation, it also causes mode collapse, highlighting a trade‑off between generalization and diversity.
By Constantinos Karouzos, Xingwei Tan, Nikolaos Aletras
arXiv:2606. 09850v1 Announce Type: new Abstract: Post-training alignment algorithms are predominantly evaluated as black boxes, obscuring how they reshape language models' internal computations.
By Aarush Sinha, Ishan Garg, Veeraraju Elluru, Arth Singh, Kushal Garg
arXiv:2509. 08022v3 Announce Type: replace-cross Abstract: Aligning large language models (LLMs) with diverse human values is essential for safe and effective deployment, yet existing benchmarks often overlook cultural and demographic variation.
By Yao Liang, Dongcheng Zhao, Feifei Zhao, Guobin Shen, Yuwei Wang, Dongqi Liang, Yi Zeng
arXiv:2509. 26169v2 Announce Type: replace Abstract: Alignment of large language models remains a central challenge in natural language processing.
By Fr\'ed\'eric Berdoz, Luca A. Lanzend\"orfer, Ren\'e Caky, Roger Wattenhofer
arXiv:2608.23149v2 Announce Type: replace-cross
Abstract: The alignment of Large Language Models heavily relies on English-centric high-quality preference data, which often leads to suboptimal perfor...
By Seungyoon Lee, Minhyuk Kim, Jungseob Lee, Heuiseok Lim
arXiv:2607. 03528v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly deployed as critical decision-making components in high-stakes real-world AI systems, rendering LLM reliability a foremost practical concern.
By Gaoxiang Luo, Yifan Wu, Sinian Zhang, Aryan Deshwal, Ju Sun
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
arXiv:2606. 03165v1 Announce Type: cross Abstract: The language used by digital chat assistants such as ChatGPT can diverge from human expectations (misalignment).
By Thomas Stephan Juzek, Xiaoyang Ming, Jose A. Hernandez