The paper introduces PLUS, a framework that uses reinforcement learning to generate text-based summaries of individual users’ preferences, characteristics, and past conversations. These summaries condition a reward model, allowing it to predict personalized response preferences and improving reward accuracy by 11–77 % over the standard Bradley‑Terry model. PLUS demonstrates robust performance with new users and topics, achieves a 25 % improvement over existing personalized RLHF techniques, and enables zero‑shot personalization for state‑of‑the‑art models like GPT‑4.
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...
arXiv:2608. 09507v1 Announce Type: cross Abstract: Natural language user preferences provide an interpretable interface for LLM personalization.
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.
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.
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.
arXiv:2606. 06614v1 Announce Type: cross Abstract: Despite growing interest, most evaluations of large language models' (LLMs') personalization abilities have relied on synthetic data.
arXiv:2605. 28882v2 Announce Type: replace-cross Abstract: With the rapid advancement of large language models, evaluating human-likeness in open-ended conversation has become increasingly important.
arXiv:2510. 05342v2 Announce Type: replace-cross Abstract: Direct Preference Optimization (DPO) has emerged as a simple and effective method for aligning large language models.
arXiv:2607. 18258v1 Announce Type: new Abstract: Reinforcement learning from human feedback (RLHF) with preference-based reward models often exhibits unstable training dynamics.
arXiv:2602. 12394v2 Announce Type: replace Abstract: Personalized prompting offers large opportunities for deploying large language models (LLMs) to diverse users, yet existing prompt optimization methods primarily focus on task-level optimization while largely overlooking user-specific preferences and latent constraints of individual users.
The paper introduces CurriPO, a tree‑structured curriculum that adapts to diverse user reward models in AI alignment. By automatically building a curriculum that branches and reuses reward models, it addresses the problem of users whose reward models are hard to optimize, a group often underserved by conventional methods. Experiments on personalized continuous control demonstrate that CurriPO improves population satisfaction by 1.2–2.1× over the best baseline while cutting training time.