1
SHRED: Retain-Set-Free Unlearning via Self-Distillation with Logit Demotion
无需保留集！新方法SHRED用自蒸馏+logit降级实现LLM高效遗忘，拒绝灾难性性能下降。
arXiv:2605.07482v2 Announce Type: replace-cross Abstract: Machine unlearning for large language models (LLMs) aims to selectively remove memorized con…