List your company

Near-duplicate identification on Full C4 (365M documents)

27,207,144Number of DuplicatesSEDD
25,231,188.7225,744,177.1126,257,165.526,770,153.89Jan 2, 2025

Evaluation results

MethodLinks
SEDD2025.0127,207,14424,533,63797
GPU baseline (NeMo Curator)2025.0125,307,187——