HarmBench
Attack simulator
A standardized framework for evaluating and improving the robustness of large language models against adversarial attacks
HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal
366 stars
6 watching
59 forks
Language: Jupyter Notebook
last commit: 6 months ago
Linked from 1 awesome list
ethicalml/awesome-production-machine-learningRelated projects:
| Repository | Description | Stars |
|---|---|---|
 robustbench/robustbench | A standardized benchmark for measuring the robustness of machine learning models against adversarial attacks | 682 |
 n0dec/malwless | A tool designed to simulate system compromise or attack behaviors without running processes or PoCs. | 271 |
 sail-research/iba | This repository provides a setup and framework for investigating irreversible backdoor attacks in Federated Learning systems. | 31 |
 markcyber/badusb | A collection of educational scripts and payloads for simulating vulnerabilities and malware attacks on Windows systems using custom hardware. | 60 |
 azure/simuland | A collaboration to create realistic test environments for simulating real-world attacks and improving detection strategies. | 704 |
 uber-common/metta | An adversarial simulation tool to test information security preparedness by simulating network-based attacks on various systems. | 1,103 |
 trycatchhcf/dumpsterfire | A toolset for creating and automating customized security events to simulate realistic scenarios for testing and training | 998 |
 borealisai/advertorch | A toolbox for researching and evaluating robustness against attacks on machine learning models | 1,311 |
 nshalabi/attack-tools | Utilities for simulating adversary behavior in the context of threat intelligence and security analysis | 1,011 |
 hfzhang31/a3fl | A framework for attacking federated learning systems with adaptive backdoor attacks | 23 |
 amv42/sshd-honeypot | An intrusion detection system designed to capture and analyze ssh interactions between an attacker and a modified OpenSSH deamon | 26 |
 splunk/attack_range | A tool to simulate attacks against virtual environments and collect data into Splunk for detection development | 2,181 |
 elastic/swat | A tool designed to simulate malicious behavior against Google Workspace environments for threat research and detection rule effectiveness testing | 163 |
 ai-secure/dba | A tool for demonstrating and analyzing attacks on federated learning systems by introducing backdoors into distributed machine learning models. | 179 |
 openbas-platform/openbas | A comprehensive cyber adversary simulation platform for planning and conducting simulated attacks and exercises | 765 |