QUAIL: Quantization Aware Unlearning for Mitigating Misinformation in LLMs
This addresses a practical issue for deploying unlearned models in resource-constrained environments, but it is incremental as it builds on existing unlearning techniques.
The paper tackles the problem of machine unlearning being undermined by low-bit quantization, which can restore forgotten information, and proposes a quantization-aware unlearning method that preserves forgetting under 4-bit quantization, outperforming existing methods.
Machine unlearning aims to remove specific knowledge (e.g., copyrighted or private data) from a trained model without full retraining. In practice, models are often quantized (e.g., 4-bit) for deployment, but we find that quantization can catastrophically restore forgotten information [1]. In this paper, we (1) analyze why low-bit quantization undermines unlearning, and (2) propose a quantization-aware unlearning method to mitigate this. We first compute weight-change statistics and bucket overlaps in quantization to show that typical unlearning updates are too small to cross quantization thresholds. Building on this insight, we introduce a logits space hinge loss: for each forget example, we force the output logits of the unlearned model to differ from the original model by at least a margin (half the quantization step). This ensures forgotten examples remain distinguishable even after quantization. We evaluate on language and classification tasks (including a Twitter misinformation dataset) and show our method preserves forgetting under 4-bit quantization, whereas existing methods almost entirely recover the forgotten knowledge.