Small single-dimension perturbations to embeddings of high-risk tokens can flip aligned LLM responses from refusal to harmful output, and a search algorithm (SEP) locates such perturbations across models.
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
CONDITIONAL 1representative citing papers
citing papers explorer
-
Embedding Poisoning: Bypassing Safety Alignment via Embedding Semantic Shift
Small single-dimension perturbations to embeddings of high-risk tokens can flip aligned LLM responses from refusal to harmful output, and a search algorithm (SEP) locates such perturbations across models.