metadata

dataset_info:
  features:
    - name: label
      dtype:
        class_label:
          names:
            '0': negative
            '1': positive
    - name: input_ids
      sequence: int32
    - name: attention_mask
      sequence: int8
  splits:
    - name: train
      num_bytes: 11392188496
      num_examples: 278021
    - name: test
      num_bytes: 2848077856
      num_examples: 69506
  download_size: 38238491
  dataset_size: 14240266352
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*
license: mit
task_categories:
  - text-classification
language:
  - ru
tags:
  - tokenized
  - toxic
  - russian
  - classify
size_categories:
  - 100K<n<1M

Tokenized Russian toxic text

Tokenized version of Mnwa/russian-toxic dataset with modernbert base model tokenizer