Liliang Ren

Baleen: Self‑Interpretable, Robust SSMs with Stochastic Selective Memory

Understanding the Mixture-of-Experts with Nadaraya-Watson Kernel

PaTH Attention: Position Encoding via Accumulating Householder Transformations

Reinforcement Learning for Reasoning in Large Language Models with One Training Example

Decoder-Hybrid-Decoder Architecture for Efficient Reasoning with Long Generation

Routing Mamba: Scaling State Space Models with Mixture-of-Experts Projection

Samba: Simple Hybrid State Space Models for Efficient Unlimited Context Language Modeling

SAS: Simulated Attention Score

S2-Attention: Hardware-Aware Context Sharding Among Attention Heads