RePolicy is a reinforcement learning approach designed to invoke safety policies for language model agents by evaluating entire execution trajectories within context-dependent policy libraries. It generates policy-grounded rationales and safety judgments, and is initialized with the PolicyTraj-20K dataset before fine-tuning via GRPO with verifiable rewards and policy-context perturbation. Experiments on six safety benchmarks demonstrate strong safety-detection performance and robust policy invocation across varying contexts.
By Houcheng Jiang, Boxuan Zhang, Qiyong Zhong, Junfeng Fang, Xiang Wang, Xiangnan He
Web agents that act from rendered pixels avoid the fragility and heavy token cost of reading a page's HTML or accessibility tree, but training them depends on large amounts of high-quality interaction...
EEG foundation models pretrained via self-supervised learning promise transferable representations, but their generalization remains limited, especially across diverse clinical datasets. Full fine-tun...
While text-based hallucination detection has been extensively studied, spoken hallucination detection remains largely unexplored, particularly for low-resource languages. We present the first multilin...
Mechanistic Localization bridges mechanistic interpretability and post-training optimization by isolating critical parameters via interpretative approaches and then guiding parameter-efficient Supervi...
With the rapid development of large-scale pre-trained language models based on Transformer architectures, their high computational and memory costs have become a major obstacle to deployment, especial...
Self-supervised pretraining enables transferable representations for medical imaging, yet most CT encoders remain biased toward coarse semantic understanding, limiting their sensitivity to fine-graine...
Search-augmented reasoning remains difficult for small language models. On-policy distillation (OPD) from trained teachers offers a promising direction, but suffers from two issues: (1) high-quality m...
Safeguarding language model agents requires assessing complete execution trajectories under context-dependent safety policies. Existing policy-aware safeguards mainly rely on prompting or supervised f...
Coding agents re-send large file reads and tool outputs to a frontier LLM every turn, and this context dominates their token bill. General-purpose prompt compressors are trained on prose and suit code...
Telephone fraud is pervasive and costly, but its inner workings are rarely observed at scale. We analyze a complete corpus of 10,211 inbound scam and spam calls -- 913 hours of audio and 330,956 trans...
Prohibitive computational and environmental costs impede the scalable deployment of Large Language Models (LLMs). Traditional compression techniques (sparsity, quantization, low-rank approximations) a...
arXiv:2602.03318v4 Announce Type: replace
Abstract: Operations Research (OR) relies on expert-driven modeling--a slow and fragile process ill-suited to novel scenarios. While large language models (L...
By Yifan Shi, Jiayi Wang, Minyi Wu, Ye Fan, Jialong Shi, Jianyong Sun
arXiv:2608.22316v1 Announce Type: cross
Abstract: Whether an intermediate stage of modern Arabic handwriting helps or hurts historical Arabic HTR is usually decided from one implementation and one co...
By Sumaih Almarshad, Maram Alamri, Dona Aloraini, Fares Altuwaim, AlJawharh AlOtaibi, Reem Alyabis, Rayah Aldawsari
arXiv:2604.19907v2 Announce Type: replace
Abstract: Recent agentic frameworks for 3D scene synthesis have advanced realism and diversity by integrating heterogeneous generation and editing tools. The...
By Yun He, Kelin Yu, Matthias Zwicker
arXiv:2608.23234v1 Announce Type: new
Abstract: In this technical report, we present a training-free framework for audio-guided video object segmentation, which integrates Multimodal Large Language M...
By Liangtao Shi, Jinxia Xie, Xiantao Hu, Ting Liu
The paper introduces a framework to tackle modality imbalance in multimodal learning by focusing on sample-level variations. It defines a Modality Gap metric to measure prediction discrepancies, models the resulting bimodal distribution with a Gaussian Mixture Model, and uses Bayesian probabilities for soft separation of balanced and imbalanced samples. A two‑stage training process—Warm‑up and Adaptive Training—reallocates loss weights based on the GMM, strengthening alignment for imbalanced samples while favoring fusion for balanced ones, and shows superior performance over existing baselines.
By Zhiwen Yu, Zhaocheng Liu, Xiaoqing Liu, Huanqiang Zeng, C. L. Philip Chen
OVIBench introduces the first standardized benchmark for evaluating vision‑language models on Online Video Question Answering under Interruption, a realistic setting where users can interrupt the model during answer generation. The benchmark categorizes interruptions into Cancellation, False Trigger, and Correction, supports both open‑ended and multiple‑choice tasks, and provides an offline simulation protocol plus a multi‑dimensional metric suite. Experiments show that OVIBench can distinguish models’ interruption‑handling abilities, particularly in following correction requests, and that fine‑tuning on the newly created OVI‑Train dataset yields significant performance gains.
By Naiming Liu, Zhiheng Wu, Shuning Wang, Tie Zhang, Bowen Liu, Tong Wang
arXiv:2608.23338v1 Announce Type: cross
Abstract: LoRA fine-tuning is standard for adapting LLMs to reranking, but it remains unclear where in the network task-specific relevance behavior is learned...
By Matthew Perlman, Atharva Nijasure, James Allan