@misc{indiciae2d1b2cf7539b,
  title = {Effective Backdoor Mitigation in Vision-Language Models Depends on the Pre-training Objective},
  author = {Sahil Verma and Gantavya Bhatt and Avi Schwarzschild and Soumye Singhal and Arnav Mohanty Das and Chirag Shah and John P Dickerson and Pin-Yu Chen and Jeff Bilmes},
  year = {2025},
  url = {https://arxiv.org/abs/2311.14948},
  note = {Source identifier: 2311.14948}
}