SKstars at SHROOM: 视觉一致性引导的零样本与LoRA适配视觉-语言模型集成
SKstars at SHROOM: Visions Agreement-Guided Ensembling of Zero-Shot and LoRA-Adapted Vision--Language Models
浏览论文内容
中文总结 AI 辅助
提出结合Qwen2.5-VL-72B零样本与LoRA适配的7B模型进行集成，用于视觉-语言模型细粒度幻觉检测，在SHROOM-Visions 2026上取得Cor+Lbl 0.2902，排名第15。
中文摘要 AI 辅助
本文描述了SKstars在SHROOM-Visions 2026上的提交，这是一个关于大型视觉-语言模型输出中细粒度幻觉检测的共享任务。该任务要求系统识别幻觉字符跨度，分配幻觉类别，并为其预测提供置信度估计。我们的方法结合了来自Qwen2.5-VL-72B-Instruct的零样本预测与LoRA适配的Qwen2.5-VL-7B-Instruct模型的预测。两个模型的输出通过一个轻量级集成过程整合，随后进行跨度细化和置信度调整。我们在一个小的内部开发子集上评估了主要系统组件，并在官方英文测试集上报告了提交系统的性能。SKstars取得了Cor+Lbl分数0.2902，在29个团队中排名第15，并分别获得了Cor和IoU分数0.3642和0.3151，在这两个指标上均排名第18。结果表明，将大型零样本模型与较小的适配模型相结合，为多语言和细粒度幻觉定位提供了一个实用框架，同时也凸显了将开发集改进转移到隐藏测试数据的难度。代码和预测：此https URL
英文摘要
This paper describes the SKstars submission to SHROOM-Visions 2026, a shared task on fine-grained hallucination detection in large vision-language model outputs. The task requires systems to identify hallucinated character spans, assign hallucination categories, and provide confidence estimates for their predictions. Our approach combines zero-shot predictions from Qwen2.5-VL-72B-Instruct with those of a LoRA-adapted Qwen2.5-VL-7B-Instruct model. The outputs of the two models are integrated through a lightweight ensemble procedure, followed by span refinement and confidence adjustment. We evaluate the main system components on a small internal development subset and report the performance of the submitted system on the official English test set. SKstars achieved a Cor+Lbl score of 0.2902, ranking 15th among 29 teams, and obtained Cor and IoU scores of 0.3642 and 0.3151, respectively, ranking 18th on both metrics. The results show that combining a large zero-shot model with a smaller adapted model provides a practical framework for multilingual and fine-grained hallucination localization, while also highlighting the difficulty of transferring development-set improvements to hidden test data. Code and predictions: https://github.com/aliathar1401/SK-Stars-shroom-visions-2026
发表机构
- Chung-Ang University（中央大学）
机构由 AI 辅助整理，请以论文原文为准。
