Erased, but Not Gone: Output Forgetting Is Not True Forgetting
For the machine unlearning community, this work reveals a fundamental gap between output-level and representation-level forgetting, challenging the validity of current evaluation practices.
The paper shows that current machine unlearning evaluations based on output-level metrics (e.g., low forget-set accuracy) can overestimate success, as models may appear forgotten at the output layer while retaining structured representation-level discrepancies relative to retraining. Empirical results across multiple methods, datasets, and models demonstrate that output forgetting does not imply true retraining-consistent forgetting.
Machine unlearning (MU) is commonly judged by output forgetting, such as low forget-set accuracy or reduced logit-level membership inference. But if output-level success can coexist with retraining-inconsistent residuals in representation space, what kind of forgetting are current evaluations actually certifying? We study this question through retraining-consistent representation forgetting, using the retrained model (i.e., trained from scratch without the forget data) as an operational reference for correct forgetting. Across multiple unlearning methods, datasets, and models, our theoretical analysis and empirical results show that standard output-level evaluation can systematically overestimate the success of unlearning. Under this stronger lens, current methods often appear forgotten at the output layer while exhibiting a structured mismatch relative to retraining. They partially align with retraining on forget samples, remain more inconsistent on retain samples, and leave residual discrepancy concentrated along retraining-related directions rather than diffuse in representation space. This structured mismatch is characterized by forget/retain asymmetry, directional mismatch, and concentrated residuals along retraining-related directions. These results suggest that current MU is often evaluated for apparent forgetting rather than retraining-consistent forgetting. More broadly, retraining reveals what output forgetting hides.