SafeVid: Toward Safety Aligned Video Large Multimodal Models | Read Paper on Bytez