arXiv:2510.17881v4 Announce Type: replace
Abstract: Large language models (LLMs) are typically aligned with population-level preferences, despite substantial variation across individual users. We int...
By Yizhuo Chen, Xin Liu, Ruijie Wang, Zheng Li, Pei Chen, Changlong Yu, Qingyu Yin, Priyanka Nigam, Meng Jiang, Bing Yin
arXiv:2608. 09507v1 Announce Type: cross Abstract: Natural language user preferences provide an interpretable interface for LLM personalization.
By Yuting Liu, Wei Wu, Jianzhe Zhao, Guibing Guo
PerSPL (Personalized Semi-supervised Preference Learning) is a backbone‑agnostic framework that stabilizes cross‑prompt score coordinates using pair‑mean score centering, enabling the conversion of unlabeled responses into user‑specific training signals. It addresses the scarcity of user‑specific preference comparisons by leveraging large pools of unlabeled responses, which are otherwise unused by existing personalized reward models. Across four benchmarks—PersonalLLM, HelpSteer2, Reddit TLDR, and Chatbot Arena—PerSPL achieves the highest mean accuracy for both seen and unseen users, demonstrating its effectiveness in summarization and open‑domain chatbot interaction settings.
By Xiaoan Xu, Tingzhou Wei, Juncheng Dong, Korel Gundem, Ethan X. Fang, Zhengling Qi
Personalized reward models require user-specific preference comparisons, but collecting such feedback is costly and provides only limited supervision for each user. At the same time, large pools of re...
Natural Language Actor-Critic (NLAC) is a new actor‑critic algorithm that trains large language model (LLM) agents by using a generative LLM critic to produce natural‑language feedback instead of scalar rewards. By providing richer, actionable explanations for suboptimal actions, NLAC helps LLM policies improve without relying on random exploration, and it can be trained off‑policy without policy gradients. Experiments on reasoning, web browsing, tool‑use, and dialogue tasks show that NLAC outperforms existing training methods and offers a more scalable and stable training paradigm for LLM agents.
By Joey Hong, Kang Liu, Zhan Ling, Jiecao Chen, Sergey Levine
The paper investigates whether using a language model’s memory to personalize responses aligns with user preferences. By collecting about 1,000 pairwise judgments from 87 participants on ChatGPT queries that included personal memories, the authors find that memory‑conditioned responses are not consistently preferred and that users sometimes strongly favor non‑memory responses. An automatic reward model trained on these judgments achieves 72.8% preference prediction accuracy and outperforms existing baselines, highlighting the need for evaluation and optimization objectives that better reflect user preferences for memory use.
By Yijiang River Dong, Zexue He, Cheng Yu, Tiancheng Hu, Michael J Ryan, Jan Batzner, Zheng Hui, Shaolong Wu, Orestis Papakyriakopoulos, Alex Pentland, Nigel Collier, Jiaxin Pei