Visual tokens enter large language models as raw, unstructured signals. Their internal transformation and integration depend on architecture—either as in-context prompts or injected into intermediate layers—leading to distinct evolution paths in visual representation and frequency characteristics. We find that attention alone is insufficient; performance is driven by the quality of visual representations at each layer across different integration paradigms.
arxiv
arXiv cs.AI
·
51d ago
·
research
Hidden Evolution of Disguised Visual Context in VLMs
from English
Related articles
arxiv
arXiv cs.LG
·
52d ago
·
3 views
OneCanvas: 3D Scene Understanding via Panoramic Reprojection
OneCanvas enables 3D scene understanding in Vision-Language Models by aggregating patch features onto a single panoramic canvas using 3D world coordinates. It achieves state-of-the-art performance on SQA3D and VSI-Bench, and generalizes to out-of-distribution data on SPBench, using significantly less training compute than existing methods.
arxiv
arXiv cs.CL
·
47d ago
·
1 view
Beaver: Agent Harness for Scientific Curation from Multimodal Sources
Beaver is an agent harness that extracts structured information from scientific papers by integrating multimodal evidence tooling, task scaffolding, and artifact-grounded autoresearch. It achieves 81.0 on the Gold-Referenced Attribute Score, outperforming frontier agents by over 23 points, with key gains on high-value attributes requiring cross-modal reasoning.
arxiv
arXiv cs.CL
·
51d ago
·
1 view
H-RePlan: Hierarchical Recovery for Cross-Device Agent Systems
H-RePlan introduces a hierarchical replanning framework that separates device-local strategy recovery from global orchestrator replanning. It outperforms existing baselines by achieving higher completion and instruction adherence, with reduced token cost, through scope-aware recovery in multi-device agent systems.
arxiv
arXiv cs.AI
·
51d ago
·
1 view
See-and-Reach: Vision-Language Navigation for UAVs in Field of View
UAV-VLN-FOV isolates the see-and-reach stage for precise evaluation of UAV navigation. 3DG-VLN enhances visual grounding and spatial alignment using dynamic 3D direction cues, achieving a 13.82% success rate improvement over baselines and validated in real-world trials.
arxiv
arXiv cs.CL
·
51d ago
·
3 views
Semantic Clusters Pre-Train Tsetlin Machine for Interpretability
A new framework pre-trains the Tsetlin Machine using semantic clusters from language models, avoiding embeddings. The method groups text samples into coherent clusters via K-means or Top2Vec, then uses cluster-sample pairs to train a non-negated TM with Type I feedback. Results show superior performance across five datasets, matching BERT-level accuracy while maintaining full interpretability.