@misc{indiciaebf7c5e796719,
  title = {Risk Awareness Injection: Calibrating Vision-Language Models for Safety without Compromising Utility},
  author = {Mengxuan Wang and Yuxin Chen and Gang Xu and Tao He and Hongjie Jiang and Ming Li},
  year = {2026},
  url = {https://arxiv.org/abs/2602.03402},
  note = {Source identifier: 2602.03402}
}