Abstract:User modeling characterizes individuals through their preferences and behavioral patterns to enable personalized simulation and generation with Large Language Models (LLMs) in contemporary approaches. However, existing methods, whether prompt-based or training-based methods, face challenges in balancing personalization quality against computational and data efficiency. We propose a novel framework CURP, which employs a bidirectional user encoder and a discrete prototype codebook to extract multi-dimensional user traits. This design enables plug-and-play personalization with a small number of trainable parameters (about 20M parameters, about 0.2\% of the total model size). Through extensive experiments on variant generation tasks, we show that CURP achieves superior performance and generalization compared to strong baselines, while offering better interpretability and scalability. The code are available at this https URL
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2602.00742 [cs.CL] |
| (or arXiv:2602.00742v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2602.00742 arXiv-issued DOI via DataCite |
Submission history
From: Liang Wang [view email]
[v1]
Sat, 31 Jan 2026 14:13:06 UTC (393 KB)
[v2]
Mon, 1 Jun 2026 14:16:38 UTC (1,146 KB)