Average segmentation accuracy and model count do not reveal whether models fail on the same cases. We present CoFailMed, an audit of correlated failures in pretrained-encoder pools for medical segmentation. Across five 2D tasks, we evaluate 11 public pretrained encoders (a controlled sample, not a census) using frozen...
Eunseob Choi, Kyeonghun Kim, Hyuk-Jae Lee et al.· 0 citations
Bridging independent vision-language models (VLMs) and small language models (SLMs) in a modular multi-agent system typically forces them to exchange information as discrete text, which creates a lossy bottleneck for spatial and fine-grained details. This constraint is especially acute for SLMs, whose limited capacity...
Mo Shahdloo, Faegheh Sardari, Mathew Salvaris· 0 citations
It is shown that on a standard evaluation dataset, simple features beat most monitors without ever running the code, outranking 20 of 24 open-weight models prompted as monitors, suggesting that a high score on such a dataset is not by itself evidence of sabotage detection.
This survey examines embodied memory as a mechanism for preserving decision-relevant information beyond the current observation, and provides a framework for understanding memory as an essential component of reliable, adaptive, and lifelong-horizon embodied intelligence.
Si-Qi Li, Zi-Yi Bai, Xinrun Xu et al.· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
OT-CAL is proposed, an adaptive data selection method that automatically adjusts its sampling preferences during training that improves performance on standard visual learning benchmarks and is three times faster training than full-data training.
Activation steering is an inference-time technique used to control or modify the behavior, tone, or safety properties of a large language model (LLM) without retraining the model's weights or editing its input prompt, but simply by adding a vector to its residual-stream activations. However, activation steering also ch...
Shivam Adarsh, Hetianyu Huang, Lau Safin et al.· 0 citations
GRIT, a training framework that incorporates visual grounding into RLVR to mitigate the impact of Correct-but-Visually-Misgrounded Reasoning (CVMR), is proposed and introduced, and experiments demonstrate that GRIT significantly improves both reasoning accuracy and grounding fidelity over existing RLVR methods.
Chanseul Cho, No-Chan Bae, Dongheon Lee et al.· 0 citations
Difficulty-Aware Chain-of-Thought (DA-CoT), a training-free adaptive framework that selectively triggers CoT based on the input's intrinsic difficulty, is proposed, finding that the entropy of the model's attention maps reliably reflects task difficulty: low entropy indicates confident, localized grounding, while high...
Chanseul Cho, Jae-Yun Jun, Dongheon Lee et al.· 0 citations
A paired protocol with empty, neutral, nonsense, infeasible, and wrong-task instructions is applied to six model checkpoints in LIBERO-Goal simulation and three methods for suppressing the start are tested.
Min-Su Jang, Jae-Woo Choi, Youngwoo Yoon et al.· 0 citations
Deploying 6G network slices across administrative domains requires an online decision for each arriving request: which domain hosts each virtual network function, how traffic crosses domain boundaries, and whether the slice is admitted. This joint decision problem is NP-hard under continuous arrivals. Experiments on a...
Mahmoud S. Mayaleh, Fetia Bannour, Nour-El-Houda Yellas· 0 citations
Long-context diffusion repeatedly pays for both token interaction and feed-forward computation. We formulate these costs as two separate budgets: the edges available to an attention graph and the experts activated for each token. This view connects local, dilated, and global attention with conditional expert computatio...
Alexandros Christoforos, Chadbourne Davis, Paul Raj· 0 citations
This work presents KEEPER, a comparative analysis of four auxiliary objectives along information-content and update-sensitivity axes within a shared recursive interface, examining how supervision targets relate to agent performance.
Jinkyeong Choi, Chaebin Jeong, Junsung Jeon et al.· 0 citations
