Zhen Wang

$\nabla$-Reasoner: LLM Reasoning via Test-Time Gradient Descent in Latent Space

DeepPersona: A Generative Engine for Scaling Deep Synthetic Personas

Learning Modal-mixed Chain-of-thought Reasoning with Latent Embedding

FIRE-Bench: Evaluating Research Agents on the Rediscovery of Scientific Insights

scPilot: Large Language Model Reasoning Toward Automated Single-Cell Analysis and Discovery

Self-MoE: Towards Compositional Large Language Models with Self-Specialized Experts