5 papers
Steering at the Source: Style Modulation Heads for Robust Persona Control
Yoshihiro Izawa, Gouki Minegishi, Koshi Eguchi +2
Activation steering offers a computationally efficient mechanism for controlling Large Language Models (LLMs) without fine-tuning. While effectively controlling target traits (e.g.…
GTaP: A GPU-Resident Fork-Join Task-Parallel System with a Pragma-Based Interface
Yuki Maeda, Kenjiro Taura
Graphics Processing Units (GPUs) excel at regular data-parallel workloads. In contrast, many irregular workloads are naturally expressed using fork-join task parallelism, which is…
Importance-Aware Data Selection for Efficient LLM Instruction Tuning
Tingyu Jiang, Shen Li, Yiyao Song +6
Instruction tuning plays a critical role in enhancing the performance and efficiency of Large Language Models (LLMs). Its success depends not only on the quality of the instruction…
How Different from the Past? Spatio-Temporal Time Series Forecasting with Self-Supervised Deviation Learning
Haotian Gao, Zheng Dong, Jiawei Yong +3
Spatio-temporal forecasting is essential for real-world applications such as traffic management and urban computing. Although recent methods have shown improved accuracy, they ofte…
Extracting Spatiotemporal Data from Gradients with Large Language Models
Lele Zheng, Yang Cao, Renhe Jiang +4
Recent works show that sensitive user data can be reconstructed from gradient updates, breaking the key privacy promise of federated learning. While success was demonstrated primar…