PPL
Stefano Ermon
Papers (5)
Direct Preference Optimization: Your Language Model is Secretly a Reward ModelFlashAttention: Fast and Memory-Efficient Exact Attention with IO-AwarenessBeyond Pairwise Preferences: Listwise Reward-Aware Alignment for Diffusion ModelsOne-Step Generative Modeling via Wasserstein Gradient FlowsGenerative Modeling with Flux Matching