Alternate Preference Optimization for Unlearning Factual Knowledge in Large Language Models
This addresses privacy and utility issues in LLMs for applications requiring data removal, though it is incremental as it builds on existing unlearning methods.
The paper tackles the problem of factual knowledge unlearning in Large Language Models (LLMs) by proposing Alternate Preference Optimization (AltPO), which combines negative and positive feedback to avoid nonsensical outputs, resulting in effective unlearning while maintaining model performance.
Machine unlearning aims to efficiently eliminate the influence of specific training data, known as the forget set, from the model. However, existing unlearning methods for Large Language Models (LLMs) face a critical challenge: they rely solely on negative feedback to suppress responses related to the forget set, which often results in nonsensical or inconsistent outputs, diminishing model utility and posing potential privacy risks. To address this limitation, we propose a novel approach called Alternate Preference Optimization (AltPO), which combines negative feedback with in-domain positive feedback on the forget set. Additionally, we introduce new evaluation metrics to assess the quality of responses related to the forget set. Extensive experiments show that our approach not only enables effective unlearning but also avoids undesirable model behaviors while maintaining overall model performance. Our implementation can be found at https://github.com/molereddy/Alternate-Preference-Optimization.