Measuring the Depth of LLM Unlearning via Activation Patching
Large language model (LLM) unlearning has emerged as a crucial post-hoc mechanism for privacy protection and AI safety, yet auditing whether target knowledge is truly erased remains challenging. Existing output-level metrics fail to detect when this knowledge remains recoverable from internal representations. Recent white-box studies reveal such residual knowledge but often rely on auxiliary training or dataset-specific adaptations, leaving no generalizable metric. We close this gap with the Unlearning Depth Score (UDS), a metric that quantifies the mechanistic depth of unlearning via activation patching. UDS first identifies layers that encode the target knowledge using a retain model ba...
| Authors | Jaeung Lee, Dohyun Kim, Jaemin Jo |
|---|---|
| Primary category | cs.CL |
| Published | 23 May 2026 |
| Last updated at source | 3 October 2026 |
Models and datasets referencing this paper
- jaeunglee/uds-annotated-tofuDataset
Silytics shows public metadata from arXiv and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.