Personalized Text Generation with Fine-Grained Linguistic Control
This work addresses the need for more nuanced personalization in text generation for users seeking tailored content, though it is incremental as it builds on existing controllable generation methods.
The paper tackles the problem of generating personalized text by controlling fine-grained linguistic attributes, such as lexical and syntactic features, and introduces a benchmark to train and evaluate models, finding that various large language models perform differently with insights into influencing factors.
As the text generation capabilities of large language models become increasingly prominent, recent studies have focused on controlling particular aspects of the generated text to make it more personalized. However, most research on controllable text generation focuses on controlling the content or modeling specific high-level/coarse-grained attributes that reflect authors' writing styles, such as formality, domain, or sentiment. In this paper, we focus on controlling fine-grained attributes spanning multiple linguistic dimensions, such as lexical and syntactic attributes. We introduce a novel benchmark to train generative models and evaluate their ability to generate personalized text based on multiple fine-grained linguistic attributes. We systematically investigate the performance of various large language models on our benchmark and draw insights from the factors that impact their performance. We make our code, data, and pretrained models publicly available.