A regime theory selects the optimal controller class for LLM action decisions from a nested lattice of four classes using three data-estimable bottlenecks, with a Bernstein-tight threshold and empirical matches on multiple benchmarks.
Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
years
2026 5representative citing papers
A new attention-enhancement method using ARS scores and RVE reduces action-relation hallucinations in LVLMs while generalizing to spatial and object hallucinations.
CAST reduces object hallucination in LVLMs by 6.03% on average across five models and five benchmarks by identifying caption-sensitive attention heads and applying optimized steering directions to their outputs, with negligible added inference cost.
ACE uses adversarial counter-commonsense perturbations on image tokens during decoding to suppress hallucinated linguistic priors while preserving stable visual signals in MLLMs.