CARD: Cluster-level Adaptation with Reward-guided Decoding for Personalized Text Generation
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 efficient fine-tuning a
Reading in Plain English mode — toggle below to switch
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 efficient fine-tuning a
This work addresses a critical challenge in LLM deployment: personalized generation at scale. By combining cluster-level adaptation with individual preference learning, it offers a practical approach to tailor models to users while maintaining efficiency, which is highly relevant for product-focused ML engineers.