Alexander Robey

Command-V: Training-Free Representation Finetuning Transfer

Jailbreaking in the Haystack

Existing Adversarial LLM Unlearning Evaluations Are Inconclusive

Benchmarking Mitigations For Covert Misuse

Towards Understanding the Transferability of Adversarial Suffixes in Large Language Models

Safety Pretraining: Toward the Next Generation of Safe AI

Antidistillation Sampling

Automated Black-box Prompt Engineering for Personalized Text-to-Image Generation

Transferable Adversarial Attack on Vision-enabled Large Language Models