{
  "version": "v1.0.1",
  "repository": "pathcosmos/korean-pretrain-training-data",
  "data_commit": "7e16a6d245fc895991c643ecfe27cf37aba6a584",
  "native_mix_sha256": "f9737e3cb2260ad0e707245ef2b69027177cdeccdc1e132dc81c19401f0e4c62",
  "native_config_sha256": "8b2f1d21902b2bc2c7d233b91f0cb607b2ae10599d262bd17eeec58ed5c4a426",
  "tokenizer_sha256": "050f9b43cf964f0c3b29ae443a1c2eae46cb2ee0022cc010bcfea3d26acf1706",
  "token_files": 732,
  "token_bytes": 242377468750,
  "consumer_validation": "passed",
  "full_3B_training_reproduction": false,
  "documentation_update": "Tokenizer loading uses an explicit dataset snapshot; reader dependencies include tokenizer packages"
}