This dataset is a recreation of new_conv_xai, after I went back and completely cleaned up the original train and test jsons.
This dataset is derived from cleaned_conv_xai. cleaned_conv_xai had the integration of all the images and other metadata together with each of the 30 conversations from the train and test json files.
cleaned_conv_xai_augmented takes it a step further and converts each of those conversations into subsets of conversation histories, creating… See the full description on the dataset page:
https://huggingface.co/datasets/mxforml/cleaned_conv_xai_augmented.
