Skip to content
Search
paperSeptember 2026Unreviewed

A Novel Semantic Manifold Alignment Attack against Embedding-to-Embedding Obfuscation in Privacy-Preserving LLMs

Sicong Li, Lingfeng Yao, Xingke Yang, Ke Tu, Chenhao Wu, Hao Wang, Jiang Liu, Phone Lin, Xin Fu, Miao Pan

View resourcePDF

Abstract

With the widespread applications of large language models (LLMs), privacy-preserving inference has become increasingly essential for sensitive queries. To balance privacy and utility, a series of lightweight obfuscation approaches has recently been proposed, where users locally transform plaintext embeddings into the fixed ciphertext ones. While such Embedding-to-Embedding Obfuscation (E2EO) schemes demonstrate considerable resilience against traditional token frequency and embedding inversion a

Categories

Guardrails

Cite

@misc{li2026novel,
  title = {{A Novel Semantic Manifold Alignment Attack against Embedding-to-Embedding Obfuscation in Privacy-Preserving LLMs}},
  author = {Sicong Li and Lingfeng Yao and Xingke Yang and Ke Tu and Chenhao Wu and Hao Wang and Jiang Liu and Phone Lin and Xin Fu and Miao Pan},
  year = {2026},
  month = sep,
  eprint = {2609.06749},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2609.06749}
}