This dataset provides a structural unlearning task for LLMs.
Recently, machine unlearning, which seeks to erase specific data stored in the pre-trained or fine-tuned models, has emerged as a crucial protective measure for LLMs. However, unlearning approaches for LLMs that have been considered thus far have focused on the removal of independent data points and have not taken into account that the stored facts are logically… See the full description on the dataset page:
https://huggingface.co/datasets/xinchiqiu/PISTOL.
