Search papers, labs, and topics across Lattice.
University of the Basque Country, Spain
4
0
6
HalluTruthQA-4K reveals that over 40% of Arabic model-generated responses contain hallucinations, with detailed annotations that pinpoint specific errors and their explanations.
No single Arabic LLM can master hallucination detection, localization, and explanation, revealing critical gaps in current models' capabilities.
TMF-RSE achieves unprecedented accuracy in lung severity scoring while quantifying uncertainty, setting a new benchmark for clinical imaging applications.
Achieve state-of-the-art medical image segmentation with a model that's both accurate and lean, needing just 10 GFLOPs and a few inference steps.
