AI Dynamics

Global AI News Aggregator

About

Low-Resource Languages ‘Jailbreak’ GPT-4, Bypassing Safety Refusals with Harmful Prompts

Low-Resource Languages Jailbreak GPT-4: Translating harmful prompts into Zulu, Scottish Gaelic, Hmong, and Guarani bypasses GPT-4 safety refusals as often as best known jailbreak prompts (79% on AdvBenchmark). Example requesting homemade bomb instructions in Scottish Gaelic:

→ View original post on X — @goodside