CARD: Cluster-level Adaptation with Reward-guided Decoding for Personalized Text Generation
On the map: LLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Abstract (arXiv)
Adapting large language models to individual users remains challenging due to the tension between fine-grained personalization and scalable deployment. We present CARD, a hierarchical framework that achieves effective personalization through progressive refinement. CARD first clusters users according to shared stylistic patterns and learns group-specific LoRA adapters, enabling robust generalization and strong low-resource performance. To capture individual differences within each cluster, we propose an implicit preference learning mechanism that contrasts user-authored text with cluster-level generations, allowing the model to infer user-specific style preferences without manual annotation. At inference time, CARD injects personalization exclusively at decoding via lightweight user preference vectors and low-rank logit corrections, while keeping the base model frozen. Experiments on the LaMP and LongLaMP benchmarks show that CARD achieves superior generation quality compared to baselines, while significantly improving efficiency and scalability for practical personalized text generation.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Placed in Human values and AI feedback alignment, one of 10 directions in LLM alignment / RLHF / preference optimisation. Builds on 34 of its 50 references that are in the map. First posted on arXiv 9 Jan 2026.
Builds on
34 of 50 references.
Closest ancestors · 6
Same direction, newest first
- 2025Personalized LLM Decoding via Contrasting Personal Preference13 Jun
- 2025Language Model Personalization via Reward Factorization8 Mar
- 2025Drift: Decoding-time Personalized Alignments with Implicit User Preferences20 Feb
- 2024Personalized Steering of Large Language Models: Versatile Steering Vectors Through Bi-directional Preference Optimization28 May
- 2024Aligning LLM Agents by Learning Latent Preference from User Edits23 Apr
- 2023Personalized Soups: Personalized Large Language Model Alignment via Post-hoc Parameter Merging17 Oct
Other references · 28
- 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,084
- 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
- 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,767
- 2019Parameter-Efficient Transfer Learning for NLPcited by 4,195
- 2021Prefix-Tuning: Optimizing Continuous Prompts for Generationcited by 3,878
- 2023Generative Agents: Interactive Simulacra of Human Behaviorcited by 3,500
- 2021Unsupervised Dense Information Retrieval with Contrastive Learningcited by 1,203
- 2022Out of One, Many: Using Language Models to Simulate Human Samplescited by 737
- 2021P-Tuning v2: Prompt Tuning Can Be Comparable to Fine-tuning Universally Across Scales and Taskscited by 686
- 2023LaMP: When Large Language Models Meet Personalizationcited by 425
- 2022Social Simulacra: Creating Populated Prototypes for Social Computing Systemscited by 300
- 2024Optimization Methods for Personalizing Large Language Models through Retrieval Augmentationcited by 98
16 more
- 2023Pearl: Personalizing Large Language Model Writing Assistants with Generation-Calibrated Retrieverscited by 91
- 2024LongLaMP: A Benchmark for Personalized Long-form Text Generationcited by 91
- 2024RAG vs Fine-tuning: Pipelines, Tradeoffs, and a Case Study on Agriculturecited by 80
- 2023Reward-Augmented Decoding: Efficient Controlled Text Generation With a Unidirectional Reward Modelcited by 79
- 2024HYDRA: Model Factorization Framework for Black-Box LLM Personalizationcited by 59
- 2024Personalized Large Language Modelscited by 44
- 2025Measuring What Makes You Unique: Difference-Aware User Modeling for Enhancing LLM Personalizationcited by 44
- 2022Towards Teachable Reasoning Systems: Using a Dynamic Memory of User Feedback for Continual System Improvementcited by 37
- 2024Personalized LLM Response Generation with Parameterized Memory Injectioncited by 33
- 2024Customizing Language Models with Instance-wise LoRA for Sequential Recommendationcited by 32
- 2025Personalized Text Generation with Contrastive Activation Steeringcited by 30
- 2025Amulet: ReAlignment During Test Time for Personalized Preference Adaptation of LLMscited by 29
- 2025PROPER: A Progressive Learning Framework for Personalized Large Language Models with Group-Level Adaptationcited by 22
- 2024Personalized LoRA for Human-Centered Text Understandingcited by 17
- 2024Lifelong Personalized Low-Rank Adaptation of Large Language Models for Recommendationcited by 16
- 2024Step-Back Profiling: Distilling User History for Personalized Scientific Writingcited by 9
Built on by
Citing papers per year: 2026: 2.
2 papers.
Similar, not cited
Close in wording, no citation either way
- 2026PrefReward: Learning User Preference Matrix for Personalized Text Generation
- 2025T-POP: Test-Time Personalization with Online Preference Feedback
- 2026Preference Heads in Large Language Models: A Mechanistic Framework for Interpretable Personalization
- 2026Rethinking Personalized Generation: Test-Time Alignment via Factorized Ranking Models
- 2026Sparse Personalized Text Generation with Multi-Trajectory Reasoning
- 2025Iterative Critique-Refine Framework for Enhancing LLM Personalization
- 2026Rethinking Personalization in Large Language Models at the Token Level
- 2026Text as a Universal Interface for Transferable Personalization