AART: AI-Assisted Red-Teaming with Diverse Data Generation for New LLM-powered Applications
This addresses the need for safe and responsible deployment of LLMs in new applications, though it is incremental as it builds on existing red-teaming efforts.
The paper tackles the problem of adversarial testing for large language models (LLMs) by introducing AART, an automated approach for generating diverse adversarial evaluation datasets, which reduces human effort and shows promising results in concept coverage and data quality.
Adversarial testing of large language models (LLMs) is crucial for their safe and responsible deployment. We introduce a novel approach for automated generation of adversarial evaluation datasets to test the safety of LLM generations on new downstream applications. We call it AI-assisted Red-Teaming (AART) - an automated alternative to current manual red-teaming efforts. AART offers a data generation and augmentation pipeline of reusable and customizable recipes that reduce human effort significantly and enable integration of adversarial testing earlier in new product development. AART generates evaluation datasets with high diversity of content characteristics critical for effective adversarial testing (e.g. sensitive and harmful concepts, specific to a wide range of cultural and geographic regions and application scenarios). The data generation is steered by AI-assisted recipes to define, scope and prioritize diversity within the application context. This feeds into a structured LLM-generation process that scales up evaluation priorities. Compared to some state-of-the-art tools, AART shows promising results in terms of concept coverage and data quality.