A Recipe For Arbitrary Text Style Transfer with Large Language Models

TL;DR

This paper introduces a zero-shot text style transfer method using large language models with a novel prompting approach called augmented zero-shot learning, enabling style transfer without fine-tuning across diverse styles.

Contribution

The paper proposes a simple prompting technique that allows large language models to perform arbitrary style transfer tasks without additional training or exemplars.

Findings

Effective on sentiment transfer

Works on diverse transformations like melodramatic or metaphor insertion

No fine-tuning required

Abstract

In this paper, we leverage large language models (LMs) to perform zero-shot text style transfer. We present a prompting method that we call augmented zero-shot learning, which frames style transfer as a sentence rewriting task and requires only a natural language instruction, without model fine-tuning or exemplars in the target style. Augmented zero-shot learning is simple and demonstrates promising results not just on standard style transfer tasks such as sentiment, but also on arbitrary transformations such as "make this melodramatic" or "insert a metaphor."