Effects of diversity incentives on sample diversity and downstream model performance in LLM-based text augmentation

TL;DR

This paper examines how different diversity incentives in prompts influence the lexical diversity of LLM-generated text and the performance of downstream models in data augmentation tasks.

Contribution

It introduces and evaluates three diversity incentive methods in LLM prompts, analyzing their impact on data diversity and downstream model accuracy across multiple models and datasets.

Findings

Taboo words increase lexical diversity most.

Hints lead to the best downstream model performance.

Chaining on previous outliers has moderate effects.

Abstract

The latest generative large language models (LLMs) have found their application in data augmentation tasks, where small numbers of text samples are LLM-paraphrased and then used to fine-tune downstream models. However, more research is needed to assess how different prompts, seed data selection strategies, filtering methods, or model settings affect the quality of paraphrased data (and downstream models). In this study, we investigate three text diversity incentive methods well established in crowdsourcing: taboo words, hints by previous outlier solutions, and chaining on previous outlier solutions. Using these incentive methods as part of instructions to LLMs augmenting text datasets, we measure their effects on generated texts lexical diversity and downstream model performance. We compare the effects over 5 different LLMs, 6 datasets and 2 downstream models. We show that diversity is most increased by taboo words, but downstream model performance is highest with hints.