Skip to content
Sagan

Paper

Residual Paving: Diagnosing the Routing Bottleneck in Selective Refusal Editing

Unreadunread

AI summary

arXiv:2605. 20262v1 Announce Type: new Abstract: We study selective refusal editing as a three-way control problem: induce non-refusal on designated edit prompts while preserving benign behavior and harmful refusals outside the edit set.