Opir: Efficient Multi-Task Safety Classification for Toxicity, Jailbreaks, Hate Speech, and Harmful Content 文章

ArXiv CS.CL2026-05-29NEWSen作者: Ihor Stepanov, Aleksandr Smechov

Opir: Efficient Multi-Task Safety Classification for Toxicity, Jailbreaks, Hate Speech, and Harmful Content · 相关人物

暂无数据