terjanq/hack-a-prompt

Tools and our test data developed for the HackAPrompt 2023 competition

37
/ 100
Emerging

This project provides tools and a collection of tested prompts designed for auditing and understanding Large Language Models (LLMs). It helps security researchers and prompt engineers explore the vulnerabilities and behaviors of LLMs by providing a framework to test various input prompts and observe the model's responses. The output includes specific prompts that can trigger unexpected or vulnerable LLM behaviors, which are valuable for understanding model safety and robustness.

No commits in the last 6 months.

Use this if you are a security researcher or an AI safety expert looking to systematically test and uncover vulnerabilities or unusual behaviors in Large Language Models using a competitive, benchmarked set of prompts and tools.

Not ideal if you are looking for a general-purpose prompt engineering library for application development or a tool to simply generate creative text, as its focus is on adversarial testing and security auditing.

LLM-security AI-safety prompt-auditing vulnerability-research AI-testing
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 8 / 25
Maturity 16 / 25
Community 13 / 25

How are scores calculated?

Stars

47

Forks

7

Language

HTML

License

MIT

Last pushed

Oct 20, 2023

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/prompt-engineering/terjanq/hack-a-prompt"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.