DAMO-NLP-SG/multilingual-safety-for-LLMs

[ICLR 2024]Data for "Multilingual Jailbreak Challenges in Large Language Models"

34
/ 100
Emerging

This project provides a unique dataset to help evaluate and improve the safety of Large Language Models (LLMs) across different languages. It offers a collection of prompts, originally in English, that have been translated into nine non-English languages to test for 'jailbreaking' — instances where LLMs bypass their safety mechanisms. Anyone responsible for deploying or ensuring the safe behavior of LLMs, such as AI safety researchers or product managers, would use this to identify vulnerabilities.

101 stars. No commits in the last 6 months.

Use this if you need to test how well your large language models resist generating unsafe content when prompted in various non-English languages.

Not ideal if you are looking for a tool to *create* multilingual content or to directly translate everyday text, as its focus is specifically on LLM safety evaluation.

AI Safety Large Language Models Multilingual AI Content Moderation Model Evaluation
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 9 / 25
Maturity 16 / 25
Community 9 / 25

How are scores calculated?

Stars

101

Forks

7

Language

License

MIT

Last pushed

Mar 07, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/transformers/DAMO-NLP-SG/multilingual-safety-for-LLMs"

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