Universal and Transferable Attacks on Aligned Language Models
-
Updated
Sep 19, 2023 - Python
Universal and Transferable Attacks on Aligned Language Models
This repo focus on how to deal with prompt injection problem faced by LLMs
This project investigates the security of large language models by performing binary classification of a set of input prompts to discover malicious prompts. Several approaches have been analyzed using classical ML algorithms, a trained LLM model, and a fine-tuned LLM model.
Vulnerable LLM Application
The Security Toolkit for LLM Interactions (TS version)
⚡ Vigil ⚡ Detect prompt injections, jailbreaks, and other potentially risky Large Language Model (LLM) inputs
LLM Security Project with Llama Guard
Evaluation of Google's Instruction Tuned Gemma-2B, an open-source Large Language Model (LLM). Aimed at understanding the breadth of the model's knowledge, its reasoning capabilities, and adherence to ethical guardrails, this project presents a systematic assessment across a diverse array of domains.
CLI tool that uses the Lakera API to perform security checks in LLM inputs
MINOTAUR: The STRONGEST Secure Prompt EVER! Prompt Security Challenge, Impossible GPT Security, Prompts Cybersecurity, Prompting Vulnerabilities, FlowGPT, Secure Prompting, Secure LLMs, Prompt Hacker, Cutting-edge Ai Security, Unbreakable GPT Agent, Anti GPT Leak, System Prompt Security.
It is a comprehensive resource hub compiling all LLM papers accepted at the International Conference on Learning Representations (ICLR) in 2024.
Example of running last_layer with FastAPI on vercel
LLM Security Platform Docs
A benchmark for evaluating the robustness of LLMs and defenses to indirect prompt injection attacks.
LLM security and privacy
Guard your LangChain applications against prompt injection with Lakera ChainGuard.
This repository contains various attack against Large Language Models.
Risks and targets for assessing LLMs & LLM vulnerabilities
User prompt attack detection system
Add a description, image, and links to the llm-security topic page so that developers can more easily learn about it.
To associate your repository with the llm-security topic, visit your repo's landing page and select "manage topics."