Pinned Repositories
openchat
OpenChat: Advancing Open-source Language Models with Imperfect Data
llm-attacks
Universal and Transferable Attacks on Aligned Language Models
JailBreak-Large-Language-Model-With-A-Malicous-System-Role
We present a novel method that can jailbreak large language model with a malicous system role. It releases the potentially unethical or illegal intention of leveraging a large language model, like ChatGPT, to breach the security measures put in place to limit its access and permissions within a controlled environment.
papers
paper学习笔记
PoisonedRAG
[USENIX Security 2025] PoisonedRAG: Knowledge Corruption Attacks to Retrieval-Augmented Generation of Large Language Models
ShiJiawenwen's Repositories
ShiJiawenwen/JailBreak-Large-Language-Model-With-A-Malicous-System-Role
We present a novel method that can jailbreak large language model with a malicous system role. It releases the potentially unethical or illegal intention of leveraging a large language model, like ChatGPT, to breach the security measures put in place to limit its access and permissions within a controlled environment.
ShiJiawenwen/papers
paper学习笔记