按年被提到
20252025 · 2 次2026
AI 安全、伦理与社会影响›模型窃取、记忆与水印·arXiv:2505.20118 · cs.CL

TrojanStego: Your Language Model Can Secretly Be A Steganographic Privacy Leaking Agent

Dominik Meier, Jan Philip Wahle, Paul Röttger, Terry Ruas, Bela Gipp · 2025
开始阅读 ↗PDFalphaXiv 讨论
被提及 · 2 人首次 2025-05-31 · 最近 2026-09-22
2不同的人
3活跃月份
3讨论串
0被引(OpenAlex)
arXiv 摘要(候选层不写导读)

As large language models (LLMs) become integrated into sensitive workflows, concerns grow over their potential to leak confidential information. We propose TrojanStego, a novel threat model in which an adversary fine-tunes an LLM to embed sensitive context information into natural-looking outputs via linguistic steganography, without requiring explicit control over inference inputs. We introduce a taxonomy outlining risk factors for compromised LLMs, and use it to evaluate the risk profile of the threat. To implement TrojanStego, we propose a practical encoding scheme based on vocabulary partitioning learnable by LLMs via fine-tuning. Experimental results show that compromised models reliably transmit 32-bit secrets with 87% accuracy on held-out prompts, reaching over 97% accuracy using majority voting across three generations. Further, they maintain high utility, can evade human detection, and preserve coherence. These results highlight a new class of LLM data exfiltration attacks that are passive, covert, practical, and dangerous.

候选层:至少 2 个不同的人、在 2 个不同月份提到过,还没到主书架的 3 人 3 个月。没有做立场标注,所以只写“被提及”。

HN 书评

推荐在前,同类按时间 · 每篇最多保留 20 条,统计用全部提及
评论Worta · 2026-09

“Related to this idea, there are interesting papers that explicitly examine the adversarial case. Basically, besides the provider hiding watermarks, one could also think of an adversary training a model to exhibit this behaviour depending on the Input of the prompt. So if you use the manipulated model, not only information about the author that the platfor…”

Spymarks, not Watermarks原评论 ↗
被提交为 story · 2 次