{"url":"/dataset/aart","name":"AART","full_name":"AI-Assisted Red-Teaming","description_markdown":"AART serves as an automated alternative to the current manual red-teaming efforts.\r\nThe primary goal is to evaluate the safety of LLM generations in various application contexts.\r\n\r\nAdversarial Testing of LLMs:\r\nAdversarial testing is crucial for ensuring the safe and responsible deployment of LLMs.\r\nThe authors introduce a novel approach to generate adversarial evaluation datasets.\r\nThese datasets are used to assess the safety of LLM outputs in real-world scenarios.\r\n\r\nKey Features of AART:\r\nDiverse Data Generation: AART generates evaluation datasets with high diversity of content characteristics. This includes concepts that are sensitive, harmful, and specific to various cultural and geographic regions and application scenarios.\r\n\r\nAI-Assisted Recipes: The data generation process is steered by AI-assisted recipes. These recipes define, scope, and prioritize diversity within the application context.\r\nStructured LLM-Generation Process: AART feeds the diverse data into a structured LLM-generation process. This helps scale up evaluation priorities.\r\nPromising Results: Compared to some state-of-the-art tools, AART demonstrates promising results in terms of concept coverage and data quality.","description_withheld":null,"homepage":"https://github.com/google-research-datasets/aart-ai-safety-dataset","introduced_date":"2023-11-14","introduced_date_note":null,"introduced_by":{"paper":"/paper/aart-ai-assisted-red-teaming-with-diverse","title":"AART: AI-Assisted Red-Teaming with Diverse Data Generation for New LLM-powered Applications","first_author":"Bhaktipriya Radharapu","url":null},"license":null,"modalities":[],"tasks":[],"languages":[],"variants":["AART"],"data_loaders":[],"num_papers_in_archive":5,"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28"},"benchmarks":[],"papers_with_a_benchmark_row":[],"syntology_totals":{"read_at":"2026-09-24T18:15:14+00:00","papers_with_samples":0,"samples_harvested":0,"samples_ran":0,"samples_unverified":0,"pointer_only_for_licence":0,"papers_with_no_sample_that_ran":0,"note":"the per-paper counts above, summed; not a rate"},"papers_note":"The archive never published its papers-using-dataset list; these are papers with a leaderboard row on this dataset's benchmarks."}