Global Clipper: Enhancing Safety and Reliability of Transformer-based Object Detection Modelsby Qutub Syed Sha, Michael…
Distributional Associations vs In-Context Reasoning: A Study of Feed-forward and Attention Layersby Lei Chen, Joan…
Quantifying Task Priority for Multi-Task Optimizationby Wooseong Jeong, Kuk-Jin YoonFirst submitted to arxiv on: 5…
Block Transformer: Global-to-Local Language Modeling for Fast Inferenceby Namgyu Ho, Sangmin Bae, Taehyeon Kim, Hyunjik…
Multi-layer Learnable Attention Mask for Multimodal Tasksby Wayner Barrios, SouYoung JinFirst submitted to arxiv on:…
Contextual Counting: A Mechanistic Study of Transformers on a Quantitative Taskby Siavash Golkar, Alberto Bietti,…
D-FaST: Cognitive Signal Decoding with Disentangled Frequency-Spatial-Temporal Attentionby Weiguo Chen, Changjian Wang, Kele Xu, Yuan…
Progressive Inference: Explaining Decoder-Only Sequence Classification Models Using Intermediate Predictionsby Sanjay Kariyappa, Freddy Lécué, Saumitra…
A Temporal Kolmogorov-Arnold Transformer for Time Series Forecastingby Remi Genet, Hugo InzirilloFirst submitted to arxiv…
Mitigate Position Bias in Large Language Models via Scaling a Single Dimensionby Yijiong Yu, Huiqiang…