---
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train.csv
dataset_info:
  features:
  - name: text
    dtype: string
  - name: label
    dtype:
      class_label:
        names:
          "0": negative
          "1": positive
---

# class-label-string-token-repro

Minimal repro dataset for class-label columns whose source CSV token is a label name instead of an integer id.