@article{defendingagainstunforeseenfailuremodes, title = {Defending Against Unforeseen Failure Modes with Latent Adversarial Training}, author = {Stephen Casper and Lennart Schulze and Oam Patel and Dylan Hadfield-Menell}, year = {2024}, eprint = {2403.05030}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2403.05030v4}, }