TinyStyler: Efficient Few-Shot Text Style Transfer with Authorship Embeddings

TL;DR

TinyStyler is a lightweight, efficient model that uses authorship embeddings and a small language model to perform few-shot text style transfer, outperforming larger models like GPT-4 in style transfer tasks.

Contribution

Introduces TinyStyler, a small and efficient model leveraging authorship embeddings for effective few-shot text style transfer, surpassing existing methods in performance.

Findings

Outperforms GPT-4 in authorship style transfer

Effective in formal-informal style transfer with human and automatic evaluations

Operates efficiently with a small 800M parameter model

Abstract

The goal of text style transfer is to transform the style of texts while preserving their original meaning, often with only a few examples of the target style. Existing style transfer methods generally rely on the few-shot capabilities of large language models or on complex controllable text generation approaches that are inefficient and underperform on fluency metrics. We introduce TinyStyler, a lightweight but effective approach, which leverages a small language model (800M params) and pre-trained authorship embeddings to perform efficient, few-shot text style transfer. We evaluate on the challenging task of authorship style transfer and find TinyStyler outperforms strong approaches such as GPT-4. We also evaluate TinyStyler's ability to perform text attribute style transfer (formal $\leftrightarrow$ informal) with automatic and human evaluations and find that the approach outperforms recent controllable text generation methods. Our model has been made publicly available at https://huggingface.co/tinystyler/tinystyler .