From the 1 of 6 linked papers with an AI index.
6 papers
How Benchmarks Mis-Score Computer-Use Agents
Zihan Dong, Zhiyuan Ma, Zekun Wang +5
The paper examines how current benchmarks for computer-use agents often give inaccurate scores due to issues in task design, trajectory observation, scoring, and reporting, and pro…
Decompose Sparsely Where You Should, Absorb Densely Where You Should No
Ruixuan Deng, Zehao Jin, Zekun Wang +1
Sparse autoencoders (SAEs) are typically trained to reconstruct the \textbf{entire} residual stream through a sparse dictionary, implicitly assuming that all activation content is…
Beyond Steering Vector: Flow-based Activation Steering for Inference-Time Intervention
Zehao Jin, Ruixuan Deng, Junran Wang +2
Activation steering has emerged as a promising alternative for controlling language-model behavior at inference time by modifying intermediate representations while keeping model p…
Sparse Feature Coactivation Reveals Causal Semantic Modules in Large Language Models
Ruixuan Deng, Xiaoyang Hu, Miles Gilberti +5
We identify semantically coherent, context-consistent network components in large language models (LLMs) using coactivation of sparse autoencoder (SAE) features collected from just…
Breaking the Treewidth Barrier in Quantum Circuit Simulation with Decision Diagrams
Bin Cheng, Ziyuan Wang, Ruixuan Deng +2
Classical simulation of quantum circuits is a critical tool for validating quantum hardware and probing the boundary between classical and quantum computational power. Existing sta…
Control Flow Adaption: An Efficient Simulation Method For Noisy Quantum Networks
Huiping Lin, Ruixuan Deng, Chris Z. Yao +2
Quantum network research at both the software stack and hardware implementation level has become an exciting area of quantum information science. Although demonstrations of small-s…