works on
black-box attack 1llm jailbreak 1moderation bypass 1prompt encoding 1security evaluation 1

From the 1 of 1 linked paper with an AI index.

collaborators
Showing cs.CRShow all

1 paper · 1 filter

cs.CR2026

RoguePrompt: Dual-Layer Encoding for Self-Reconstruction to Circumvent LLM Moderation

Benyamin Tafreshian, Prathamesh Dhake

The paper presents RoguePrompt, a method that splits a forbidden request and hides it using two simple ciphers (Vigenere and ROT13) plus natural‑language reconstruction instruction…