Preprint
Aug 2026
Distance Is Not Enough: Forget-Retain Alignment Gap Predicts LLM Relearning Robustness
The Forget-Retain Alignment Gap is introduced, a training-free predictor that scores an update's forget-retain alignment without running a relearning attack, and separates selective from dense updates more reliably than global distance, suggesting that weight selectivity better explains robustness than distance alone.
Yi Chen, Hanna Hsieh, Shu-Hong Liu et al.
· 0 citations