Jiawei Zhou

Assistant Professor@State University of New York at Stony Brook·美国·OpenReview

研究方向

multimodal LMs · faithful generation · retrieval augmentation · multimodal models · LLM inference efficiency · decoding · large language models · prompting · reasoning · mixture of expert models · MoE · simultaneous machine translation · semantic parsing · structured prediction · graph neural networks · non-autoregressive neural machine translation · non-autoregressive generation · deep learning · Transformer models

Jiawei Zhou

Distilling to Hybrid Attention Models via KL-Guided Layer Selection

On the Predictive Power of Representation Dispersion in Language Models

Hidden Markov Modeling of Reasoning Dynamics in Large Language Models

Features Emerge as Discrete States: The First Application of SAEs to 3D Representations

Substituting from the Input: Distilling Sequential Computation in Transformer Language Models

Improving Chain-of-Thought Efficiency for Autoregressive Image Generation

Reduce What You Use: Input‑Aware Matrix‑Multiplication Pruning for LLMs

Safeguarding Vision-Language Models via Dynamic Token Reweighting

OKBench: Democratizing LLM Evaluation with Fully Automated, On-Demand Open Knowledge Benchmarking

LDC: Learning to Generate Research Idea with Dynamic Control

Beyond Expectations: Quantile-Guided Alignment for Risk-Calibrated Language Models

Chunk-Distilled Language Modeling

MJ-Bench: Is Your Multimodal Reward Model Really a Good Judge for Text-to-Image Generation?

Enhancing Vision-Language Model Reliability with Uncertainty-Guided Dropout Decoding

Demystifying the Underappreciated Long-Tail Problems in Large Vision Language Models