eeking arXiv cs.LG endorsement for a paper on activation probe transfer failure for reward hacking detection. I test whether probes trained on the School of Reward Hacks dataset (Taylor et al. 2025) transfer to GRPO-induced reward seeking. They don’t. The SFT and RL probe directions are nearly orthogonal (cosine = -0.07). Paper builds on Wilhelm et al. 2026, Taufeeque et al. 2026, and Gupta & Jenner 2025 (NeurIPS MechInterp Workshop). Paper will be visible on arXiv once endorsed […]

Ver mais

Federated active learning (FAL) seeks to reduce annotation cost under privacy constraints, yet its effectiveness degrades in realistic settings with severe global class imbalance and highly heterogeneous clients. We conduct a systematic study of query-model selection in FAL and uncover a central insight: the model that achieves more class-balanced sampling, especially for minority classes, consistently leads to better final performance. Moreover, global-model querying is beneficial only when the global distribution is highly imbalanced and client data are relatively […]

Ver mais

arXiv:2601.12209v1 Announce Type: new Abstract: The Fast Fourier Transform (FFT) is a fundamental numerical technique with widespread application in a range of scientific problems. As scientific simulations attempt to exploit exascale systems, there has been a growing demand for distributed FFT algorithms that can effectively utilize modern heterogeneous high-performance computing (HPC) systems. Conventional FFT algorithms commonly encounter performance bottlenecks, especially when run on heterogeneous platforms. Most distributed FFT approaches rely on static task distribution and require synchronization barriers, […]

Ver mais

Instead of catching you off-guard with a jump scare this Halloween season, EFF is here to catch you up on the latest digital rights news with our EFFector newsletter! In this issue, we’re helping you take control of your online privacy with Opt Out October; explaining the UK’s attack on encryption and why it’s bad for all users; and covering shocking new details about an abortion surveillance case in Texas. Prefer to listen in? Check out our audio […]

Ver mais

arXiv:2603.20220v1 Announce Type: new Abstract: The integration of artificial intelligence (AI) in higher education underscores the growing importance of faculty AI literacy and competency across teaching, research, and service. Existing AI literacy instruments, however, primarily target the general public, students, or K-12 teachers, and therefore lack the role-embedded indicators and psychometric validation needed for scalable assessment among university faculty. Grounded in the Critical Tech-resilient Literacies (CTRL) framework, this study develops and validates the Faculty Artificial Intelligence Literacy and […]

Ver mais

arXiv:2602.09169v1 Announce Type: new Abstract: Fully finetuning foundation language models (LMs) with billions of parameters is often impractical due to high computational costs, memory requirements, and the risk of overfitting. Although methods like low-rank adapters help address these challenges by adding small trainable modules to the frozen LM, they also increase memory usage and do not reduce inference latency. We uncover an intriguing phenomenon: sparsifying specific model rows and columns enables efficient task adaptation without requiring weight tuning. […]

Ver mais

Heterophily is a prevalent property of real-world graphs and is well known to impair the performance of homophilic Graph Neural Networks (GNNs). Prior work has attempted to adapt GNNs to heterophilic graphs through non-local neighbor extension or architecture refinement. However, the fundamental reasons behind misclassifications remain poorly understood. In this work, we take a novel perspective by examining recurring inductive subgraphs, empirically and theoretically showing that they act as spurious shortcuts that mislead GNNs and reinforce non-causal correlations […]

Ver mais

arXiv:2601.09713v1 Announce Type: new Abstract: Proactively predicting a users next utterance in human-machine dialogue can streamline interaction and improve user experience. Existing commercial API-based solutions are subject to privacy concerns while deploying general-purpose LLMs locally remains computationally expensive. As such, training a compact, task-specific LLM provides a practical alternative. Although user simulator methods can predict a user’s next utterance, they mainly imitate their speaking style rather than advancing the dialogue. Preference data synthesis has been investigated to generate […]

Ver mais