Brandon Amos

Cultivating Pluralism In Algorithmic Monoculture: The Community Alignment Dataset

BaNEL: Exploration Posteriors for Generative Modeling Using Only Negative Rewards

Wasserstein Flow Matching: Generative Modeling Over Families of Distributions

Adjoint Sampling: Highly Scalable Diffusion Samplers via Adjoint Matching

AdvPrefix: An Objective for Nuanced LLM Jailbreaks

Wasserstein Flow Matching: Generative modeling over families of distributions

Exact Byte-Level Probabilities from Tokenized Language Models for FIM-Tasks and Model Ensembles

Meta Flow Matching: Integrating Vector Fields on the Wasserstein Manifold

AdvPrompter: Fast Adaptive Adversarial Prompting for LLMs

AdvPrompter: Fast Adaptive Adversarial Prompting for LLMs

Online Intrinsic Rewards for Decision Making Agents from Large Language Model Feedback