train/mlfoundations-dclm-baseline-1.0-parquet-sampled-v3/split_10002-tokenized-chunked-1024-512-128-backfill-nodups/num_tokens.json

{
  "num_tokens": 38012102,
  "num_truncated_tokens": 37981996
}