Shangbin Feng

PhD student@University of Washington·美国·OpenReview

研究方向

natural language processing · knowledge bases · social network analysis · graph neural networks

5.5

Don't Throw Away Your Pretrained Model

ICLR 2026Poster

一作

5.5

MMMG: A Comprehensive and Reliable Benchmark for Multitask Multimodal Generation

ICLR 2026Rejected

4.5

Heterogeneous Swarms: Jointly Optimizing Model Roles and Weights for Multi-LLM Systems

NeurIPS 2025Poster

一作

6.8

Model Swarms: Collaborative Search to Adapt LLM Experts via Swarm Intelligence

ICLR 2025Rejected

一作

6.5

Varying Shades of Wrong: Aligning LLMs with Wrong Answers Only

合作者 (20)

Vidhisha Balachandran

Shangbin Feng

Don't Throw Away Your Pretrained Model

MMMG: A Comprehensive and Reliable Benchmark for Multitask Multimodal Generation

GuessBench: Sensemaking Multimodal Creativity in the Wild

MARVEL: Modular Abstention for Reliable and Versatile Expert LLMs

Generalizable LLM Learning of Graph Synthetic Data with Post-training Alignment

Model Swarms: Collaborative Search to Adapt LLM Experts via Swarm Intelligence

Sparta Alignment: Collectively Aligning Multiple Language Models through Combat

Heterogeneous Swarms: Jointly Optimizing Model Roles and Weights for Multi-LLM Systems

Model Swarms: Collaborative Search to Adapt LLM Experts via Swarm Intelligence

Varying Shades of Wrong: Aligning LLMs with Wrong Answers Only