Activation steering on probe-selected attention heads flips LLM privacy refusals into disclosures, with claims of high rates of factually correct personal information leakage.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
PII Jailbreaking in LLMs via Activation Steering Reveals Personal Information Leakage
Activation steering on probe-selected attention heads flips LLM privacy refusals into disclosures, with claims of high rates of factually correct personal information leakage.