Geometric Unlearning distills a low-rank safe subspace from reference prompts and applies projection-based alignment on synthetic anchors to suppress target content while preserving non-target utility.
Knowledge Unlearning for Mitigating Privacy Risks in Language Models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
No citing papers match the current filters.