AI RESEARCH
TEAR: Temporal-aware Automated Red-teaming for Text-to-Video Models
arXiv CS.CV
•
ArXi:2511.21145v2 Announce Type: replace Text-to-Video (T2V) models are capable of synthesizing high-quality, temporally coherent dynamic video content, but the diverse generation also inherently