---
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
dataset_info:
  features:
  - name: conversation
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  splits:
  - name: train
    num_bytes: 1586702360
    num_examples: 891857
  download_size: 860199038
  dataset_size: 1586702360
license: apache-2.0
task_categories:
- text-generation
language:
- en
pretty_name: Dolphin
---
# Dataset Card for "Dolphin"

This Dataset is edited version of [cognitivecomputations/dolphin](https://huggingface.co/datasets/cognitivecomputations/dolphin) which 
only contains conversations and GPT-4 Responses to make it easier to use it with SFTTrinaer