Genghan Zhang

Mixture-of-Mamba: Enhancing Multi-Modal State-Space Models with Modality-Aware Sparsity

Learning to (Learn at Test Time): RNNs with Expressive Hidden States

Mixture of Attention Spans: Optimizing LLM Inference Efficiency with Heterogeneous Sliding-Window Lengths

Adaptive Self-improvement LLM Agentic System for ML Library Development

Learning to (Learn at Test Time): RNNs with Expressive Hidden States

MoA: Mixture of Sparse Attention for Automatic Large Language Model Compression