metadata
dataset_info:
features:
- name: output
dtype: string
- name: in
dtype: string
splits:
- name: train
num_bytes: 37224368
num_examples: 54568
download_size: 21807129
dataset_size: 37224368
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
This is simply the teknium/GPT4-LLM-Cleaned dataset with its instruction and input columns combined into a single "in" column. I did this for simplicity, hence the name, and also because i couldnt find a similar dataset. Hope this helps!