Unveiling and Eliminating the Shortcut Learning for Locate-Then-Edit Knowledge Editing via Both Subject and Relation Awareness
This addresses a key challenge in making knowledge editing more precise and controllable for AI safety and model maintenance, though it is incremental as it builds on existing locate-then-edit frameworks.
The paper tackles the problem of uncontrollable knowledge editing in large language models, where existing locate-then-edit methods inadvertently modify unrelated relations due to shortcut learning, and proposes a two-stage optimization process that balances subject and relation features to achieve optimal overall performance.
Knowledge editing aims to alternate the target knowledge predicted by large language models while ensuring the least side effects on unrelated knowledge. An effective way to achieve knowledge editing is to identify pivotal parameters for predicting factual associations and modify them with an optimization process to update the predictions. However, these locate-then-edit methods are uncontrollable since they tend to modify most unrelated relations connected to the subject of target editing. We unveil that this failure of controllable editing is due to a shortcut learning issue during the optimization process. Specifically, we discover two crucial features that are the subject feature and the relation feature for models to learn during optimization, but the current optimization process tends to over-learning the subject feature while neglecting the relation feature. To eliminate this shortcut learning of the subject feature, we propose a novel two-stage optimization process that balances the learning of the subject feature and the relation feature. Experimental results demonstrate that our approach successfully prevents knowledge editing from shortcut learning and achieves the optimal overall performance, contributing to controllable knowledge editing.