3 papers
cs.LG2026
Representation as a Bottleneck for Mechanistic Interpretability: The Manifestation Unit Protocol
Hussein Chouman, Wataru Sasaki, Tomokazu Matsui +2
Mechanistic interpretability has produced a rich inventory of component-level analyses that characterise what neural-network components encode and how they interact. Their outputs,…
cs.LG2026
Enabling Vibration-Based Gesture Recognition on Everyday Furniture via Energy-Efficient FPGA Implementation of 1D Convolutional Networks
Koki Shibata, Tianheng Ling, Chao Qian +4
The growing demand for smart home interfaces has increased interest in non-intrusive sensing methods like vibration-based gesture recognition. While prior studies demonstrated feas…
cs.CL2025
Merging Continual Pretraining Models for Domain-Specialized LLMs: A Case Study in Finance
Kentaro Ueda, François Portet, Hirohiko Suwa +1
While LLMs excel at general tasks, they struggle in specialized domains like finance, requiring diverse skills in domain knowledge, mathematical reasoning, and multilingual process…