Faithful Dual-constrained Erasure for Robust LLM Safety Alignment
FDCU uses dual-constrained projection to erase hazardous LLM knowledge and resist retraining attacks that revive unlearned behavior.
1 stories, newest first · one card per story
FDCU uses dual-constrained projection to erase hazardous LLM knowledge and resist retraining attacks that revive unlearned behavior.
Today 556 readers · 634 views · 207 reads · all time 187K views