Yandong Wen

Streaming Autoregressive Video Generation via Diagonal Distillation

FormalMATH: Benchmarking Formal Mathematical Reasoning of Large Language Models

Stable Preference Optimization via Two-sided Contrastive Learning

Model Merging with Functional Dual Anchors

SimKO: Simple Pass@K Policy Optimization

Rethinking Preference Alignment for Diffusion Models with Classifier-Free Guidance

Advancing Formal Mathematical Reasoning with Explorative Reinforcement Learning

Parameter-Efficient Orthogonal Finetuning via Butterfly Factorization