---
dataset_info:
  features:
  - name: category
    dtype: string
  - name: conversation
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  splits:
  - name: train
    num_bytes: 2929351.0
    num_examples: 1030
  download_size: 1681823
  dataset_size: 2929351.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---

This dataset is a copy of [GAIR/lima](https://huggingface.co/datasets/GAIR/lima) transformed to work with hugging face chat templates.