beyond_the_lab_neurips_paper

license: cc-by-nc-4.0 tags: - AI-generated visual content detection - human-labeled dataset - multi-signal evaluation pretty_name: >- An anonymized eval dataset with multi-rater annotations to support reproducible research on production-scale AI content detection. task_categories: - image-classification size_categories: - 100K<n<1M notes: - The dataset is intended for benchmarking purposes only. - To reproduce the results of the paper run df = pd.read_csv("data/data.csv") print(single_method_stats(df)) print(combined_method_stats(df)) ---