research
Distance Failed to Predict What the Model Would Relearn
FRAG scores whether an unlearning update targets forget-critical weights while sparing retain-critical ones.
Summary
FRAG scores whether an unlearning update targets forget-critical weights while sparing retain-critical ones.
The authors argue that global weight displacement confuses selective unlearning with random or destructive change. Their training-free Forget-Retain Alignment Gap better separated selective from dense updates, and a pruning method built on the same principle improved relearning robustness in the reported experiments.
Why it matters
FRAG scores whether an unlearning update targets forget-critical weights while sparing retain-critical ones.
Limits and context
No additional limitation was separately recorded.
Key claims
FRAG scores whether an unlearning update targets forget-critical weights while sparing retain-critical ones.
Evidence: source-2026-08-27-009
Sources
- arXiv preprint 2608.25429arXiv · primary research
Corrections
No corrections have been recorded for this story.