HKU-TASR / ImperioLinks
[IJCAI 2024] Imperio is an LLM-powered backdoor attack. It allows the adversary to issue language-guided instructions to control the victim model's prediction for arbitrary targets.
☆42Updated 3 months ago
Alternatives and similar repositories for Imperio
Users that are interested in Imperio are comparing it to the libraries listed below
Sorting:
- Implementation of BEAST adversarial attack for language models (ICML 2024)☆87Updated last year
- General research for Dreadnode☆23Updated 11 months ago
- Adversarial Tokenization☆22Updated last month
- Tree of Attacks (TAP) Jailbreaking Implementation☆109Updated last year
- ☆40Updated 8 months ago
- Stage 1: Sensitive Email/Chat Classification for Adversary Agent Emulation (espionage). This project is meant to extend Red Reaper v1 whi…☆41Updated 9 months ago
- ☆28Updated last year
- A folder to serve tools from during PT/Red Team engagements. Contains common executables and scripts for privesc, recon, pivoting and CVE…☆14Updated last year
- The official implementation of our pre-print paper "Automatic and Universal Prompt Injection Attacks against Large Language Models".☆48Updated 7 months ago
- Enumerate SSN (System Service Numbers or Syscall ID) and syscall instruction address in ntdll module by parsing the PEB of the current pr…☆21Updated last year
- It's a handy tool to help you analyze malware. You can use this tool to query your malware samples using different hashes or find all oth…☆15Updated 2 years ago
- Exploits with pwntools library in Python3. ROP, BOF, SHELLCODE.☆21Updated last year
- ☆28Updated last year
- Mass Assigner is a simple tool made to probe for mass assignment vulnerability through JSON field modification in HTTP requests☆12Updated 11 months ago
- Identify binaries with Authenticode digital signatures signed to an internal CA/domain☆40Updated last year
- A C# Tool to gather information about email breaches☆14Updated last year
- A tool to dump users's .plist on a Mac OS system and to convert them into a crackable hash☆51Updated 7 months ago
- This library provides functionality for fetching and parsing information about Common Vulnerabilities and Exposures (CVE) using the NIST …☆14Updated 2 years ago
- using ML models for red teaming☆43Updated last year
- An improvement and a different approach to Mockingjay Self-Injection.☆34Updated last year
- A compact toolbox for backdoor attacks and defenses.☆173Updated 10 months ago
- RepoReaper is an automated tool crafted to meticulously scan and identify exposed .git repositories within specified domains and their su…☆34Updated last year
- ☆14Updated 5 months ago
- This is the ringzer0 writeup of web exploitation catagery. The name is "Word mean something"☆14Updated last year
- PAL: Proxy-Guided Black-Box Attack on Large Language Models☆51Updated 9 months ago
- ☆44Updated 2 years ago
- [EMNLP 2024] Holistic Automated Red Teaming for Large Language Models through Top-Down Test Case Generation and Multi-turn Interaction☆14Updated 6 months ago
- Source code of "TRAP: Targeted Random Adversarial Prompt Honeypot for Black-Box Identification", ACL2024 (findings)☆11Updated 6 months ago
- This repository provides a benchmark for prompt Injection attacks and defenses☆216Updated this week
- A BrainF*ck Inspired Shell Obfuscation Proof-of-Concept☆16Updated last year