SOURCE-LINKED INTELLIGENCE

Overview of SHROOM-Visions 2026: A Shared Task on Hallucination Detection in Large Vision-Language Models

arXiv · AI, language, vision and robotics · article · Aug 26, 2026 · UTC

In 2026, we held the fourth iteration of the SHROOM Shared Task series: SHROOM-Visions (\textbf{S}hared-task on \textbf{H}allucinations and \textbf{R}elated \textbf{O}bservable \textbf{O}vergeneration \textbf{M}istakes in \textbf{Vision} language model\textbf{s}), which is hosted at the UncertaiNLP Workshop co-located with EMNLP 2026. Following the success of the 2024 and 2025 tasks, this time we aim to tackle hallucinations through a model-agnostic detection task focused on large vision-language models. Building on the recently introduced SHEEP dataset, designed for long-term evaluation acros

Read original source ↗ Open in workspace

recordType

paper

region

Global