ToxiGAN: Toxic Data Augmentation via LLM-Guided Directional Adversarial Generation

TL;DR

ToxiGAN is a novel framework that uses large language models and adversarial training to generate diverse, class-specific toxic language data, improving toxicity classifier robustness.

Contribution

It introduces a dynamic, LLM-guided adversarial augmentation method with semantic ballast and directional training to enhance toxic data generation.

Findings

Outperforms existing augmentation methods on hate speech benchmarks.

Improves macro-F1 and hate-F1 scores significantly.

Semantic ballast and directional training enhance robustness.

Abstract

Augmenting toxic language data in a controllable and class-specific manner is crucial for improving robustness in toxicity classification, yet remains challenging due to limited supervision and distributional skew. We propose ToxiGAN, a class-aware text augmentation framework that combines adversarial generation with semantic guidance from large language models (LLMs). To address common issues in GAN-based augmentation such as mode collapse and semantic drift, ToxiGAN introduces a two-step directional training strategy and leverages LLM-generated neutral texts as semantic ballast. Unlike prior work that treats LLMs as static generators, our approach dynamically selects neutral exemplars to provide balanced guidance. Toxic samples are explicitly optimized to diverge from these exemplars, reinforcing class-specific contrastive signals. Experiments on four hate speech benchmarks show that ToxiGAN achieves the strongest average performance in both macro-F1 and hate-F1, consistently outperforming traditional and LLM-based augmentation methods. Ablation and sensitivity analyses further confirm the benefits of semantic ballast and directional training in enhancing classifier robustness.