In-Context Unlearning: Language Models as Few Shot Unlearners
This addresses the challenge of efficient data removal in LLMs for compliance with regulations like the Right to be Forgotten, offering a practical solution without computational overhead.
The paper tackles the problem of machine unlearning for large language models (LLMs) by proposing in-context unlearning, which removes the impact of specific training instances without updating model parameters. The result shows it performs on par or better than state-of-the-art methods that require parameter access, effectively removing instance influence while preserving test accuracy.
Machine unlearning, the study of efficiently removing the impact of specific training instances on a model, has garnered increased attention in recent years due to regulatory guidelines such as the \emph{Right to be Forgotten}. Achieving precise unlearning typically involves fully retraining the model and is computationally infeasible in case of very large models such as Large Language Models (LLMs). To this end, recent work has proposed several algorithms which approximate the removal of training data without retraining the model. These algorithms crucially rely on access to the model parameters in order to update them, an assumption that may not hold in practice due to computational constraints or having only query access to the LLMs. In this work, we propose a new class of unlearning methods for LLMs called ``In-Context Unlearning.'' This method unlearns instances from the model by simply providing specific kinds of inputs in context, without the need to update model parameters. To unlearn specific training instances, we present these instances to the LLMs at inference time along with labels that differ from their ground truth. Our experimental results demonstrate that in-context unlearning performs on par with, or in some cases outperforms other state-of-the-art methods that require access to model parameters, effectively removing the influence of specific instances on the model while preserving test accuracy.