This dataset is useful for training large language models.
The train split provides instruction-following and chat data for supervised fine-tuning (SFT) and instruction tuning.
The pretrain split… See the full description on the dataset page: https://huggingface.co/datasets/lenamerkli/distilled-web.
