This repo covers implementations of O3 framework that includes an Orthogonal low-rank adapter (LoRA) for continually unlearning requested data and an Out-Of-Distribution (OOD) detector to measure the similarity between input and unlearning data in On Large Language Model Continual Unlearning. The paper is accepted to ICLR 2025.
Parts of our code are inspired by the implementations of SOUL and ScienceQA project. We thank the authors of these repositories for their valuable work.
We provided the code for ScienceQA experiments and an initial version of our code on TOFU
-
ScienceQA_experiment -
TOFU_experiment
We will upload the clear version of the code.