While existing data attribution methods can identify which training examples build specific mechanistic circuits, they cannot explain how training data shapes the behavioral decisions a model learns. We introduce Symbolic Mechanistic Data Attribution (SMDA), a framework that attributes training pairs to the interpretab...
Probe-based uncertainty estimation (UE) has emerged as a prominent approach to detect hallucinations in Large Language Models (LLMs) by learning uncertainty from internal model signals. Yet, recent methods vary simultaneously across feature design, training data construction, and evaluation setting, obscuring what actu...
Ponhvoan Srey, Xiaobao Wu, Cong-Duy Nguyen et al.· 0 citations
This work proposes a test-time adaptation (TTA) approach that adapts to distribution shifts by leveraging homogeneity among unlabeled samples observed at inference time and establishes that TTA is a promising framework for AI text detection in the wild.
Ke Ren, Manish Raghavan, Nikhil Garg· arXiv.org· 1 citation· ⚡1
Multimodal Large Language Model (MLLM)-based web agents provide practical, high-precision solutions for visual browser automation; however, they inherently expand the attack surface, introducing novel vision-based vulnerabilities. Existing adversarial evaluations targeting these agents frequently rely on permissive thr...
Xuelong Dai, Jianyu Ma, Boyang Ma et al.· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
Evaluation of VOiLA on three benchmark problems indicates that VOiLA achieves equal or better performance than current reinforcement learning methods for POMDPs, while using less than 10% of the training data, and generalizes much better to unseen environment configurations.
Marcus Hoerger, Rishikesh Joshi, Rahul Shome et al.· arXiv.org· 0 citations
ANEForge is a Python package that programs the Apple Neural Engine (ANE), the fixed-function neural accelerator on every recent Apple device, directly and without CoreML. In production the engine is reachable only through CoreML, which treats it as a scheduling option: no configuration requires the ANE, and a model can...
Post-training LLM pruning relies on a small calibration set, usually chosen heuristically and judged by WikiText perplexity or by mean accuracy on general-domain commonsense benchmarks. This single-axis view hides capability-specific effects: at 60\% sparsity, strategies separated by $2.85$ points in averaged commonsen...
Hu Xu, Zhaolong Xing, Congcong Liu et al.· 0 citations
Video multimodal large language models (MLLMs) have made rapid progress in general and long-form video understanding, yet their ability to capture and use brief, answer-critical visual evidence remains underexplored. Momentary actions or state transitions may determine an answer while occupying only a small fraction of...
Xiaolin Liu, Yilun Zhu, Xiangyu Zhao et al.· 0 citations
The pretraining data mixture of Large Language Models (LLMs) constitutes their "digital DNA", shaping model behaviors, capabilities, and failure modes. Yet this composition is rarely disclosed, making post-hoc auditing of data combination or provenance difficult. In this work, we formalize Data Mixture Surgery (DMS): g...
Yaxin Luo, Jiacheng Cui, Xiaohan Zhao et al.· 0 citations
Standard accuracy on binary reasoning benchmarks hides critical failure modes: prior collapse, inconsistency under paraphrase, and inability to reason about parameter-dependent dynamics. We present ChaosBench-Logic v2, a 40,886-question benchmark over 165 dynamical systems with 27 FOL predicates and 78 axiom edges, tog...
EEG-based visual decoding aims to establish a mapping between neural signals and visual semantics. However, it remains constrained by the dual challenges of severe information granularity mismatch and the low signal-to-noise ratio (SNR) of EEG signals. Existing approaches typically treat visual features as static, igno...
Fan Yin, Chuhang Zheng, Peiliang Gong et al.· 0 citations
Language-based agent control is introduced, a new programming model for agentic applications that brings techniques from programming languages and language-based security to the problem of agent control by requiring agents to generate programs that are themselves well typed in the context of the surrounding scaffolding...
