arXiv cs.CL· Xiaoyu Ma, Zheyuan Zhang, Kaiwen Shi, Yanfang Ye·· 13 小时前AI 评分22
超越正确答案：组合线索如何影响视觉判断
Beyond Correct Answers: How Compositional Cues Shape Visual Judgments
AI 导读
研究通过固定图像与陈述、隔离字段标签效应，在十个模型和六项任务上考察结构化提示中各标签对视觉绑定判断的作用。结果显示，存在关系标签时加入对象角色标签对 Qwen3 在 Spatial 任务上有帮助，却损害 Qwen2；Qwen3 在保留提示下的优势主要来自交换提示引发的错误，保留提示本身接近基线。
正文 · 原文
Abstract:Visual question answering requires models to assign attributes and relations to the correct objects, a problem known as visual binding. Structured prompts make these assignments explicit through labeled fields. Yet an improvement from the full prompt does not reveal which labels help or how they interact. First, we isolate field-label effects across ten models and six tasks by holding the image and claim fixed. These comparisons reveal differences that overall accuracy can hide. With a relation label present, adding object-role labels helps Qwen3 but hurts Qwen2 on Spatial. Second, we compare auxiliary hints that preserve or swap the question's role assignments against a no-hint baseline. Across two image sets, Qwen3's advantage with preserving hints mainly reflects errors caused by swapped hints. Preserving hints remain near baseline. Field order further affects the harm. Third, we test clearer hint wording and direct selection of the relation shown in the image. Clearer hints improve text matching but do not consistently improve visual judgments. In contrast, relation selection partly reduces errors for some hints in Qwen3 and GPT-4.1-mini.
| Comments: | 29 pages, 6 figures |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.28843 [cs.CV] |
| (or arXiv:2610.28843v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.28843 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Xiaoyu Ma [view email]
[v1]
Tue, 20 Oct 2026 19:16:00 UTC (21,395 KB)
来源：arXiv cs.CL · arxiv.org