arXiv cs.AI· Jiahao Huang, Kon Woo Kim, Song-Duo Ma, Yuka Kitamura, Felix Wagner, Yunshan Huang, Sunisth Kumar, Xanh Ho, Yun-Ang Wu, Zhijie Chen, Yunjia Xu, Junfeng Jiang, Zefan Yu, Fei Cheng, Akiko Aizawa·· 3 小时前AI 评分46
契诃夫之枪为何总开火：文生图模型中的过度实现问题
When Chekhov's Gun Fires Anyway: Over-Realization in Text-to-Image Models
AI 导读
arXiv 论文提出文生图生成中的“过度实现”失败：提示词提及但未要求出现的目标出现在生成图像中，涵盖否定、人物想象、比喻和视角遮挡四类情境。团队构建 OverReal 基准，含 698 条提示词和 2,841 张人工标注图像，评估 8 个文生图系统发现均存在此问题，而 CLIPScore 等常规指标反而奖励该失败；提示词扩展只能部分缓解。
正文 · 原文
Authors:Jiahao Huang, Kon Woo Kim, Song-Duo Ma, Yuka Kitamura, Felix Wagner, Yunshan Huang, Sunisth Kumar, Xanh Ho, Yun-Ang Wu, Zhijie Chen, Yunjia Xu, Junfeng Jiang, Zefan Yu, Fei Cheng, Akiko Aizawa
Abstract:In text-to-image generation, a target that the prompt mentions often appears in the generated image in a way the prompt does not call for. We call this failure over-realization. Mental-spaces theory points to the situations where it can happen: under a negation, inside what a character believes or imagines, as the vehicle of a figurative comparison, or behind a viewpoint that hides the target. Prior work has studied only negation in depth, and the conventional alignment and preference metrics do not penalize the failure. We build OverReal to measure it: OverReal-Gen, 698 prompts covering the four situations, on which we measure each generator's over-realization rate; and OverReal-Det, 2,841 generated images human-labeled for whether and how the target is over-realized, on which we verify the automatic judge that produces this rate. Our evaluation of eight text-to-image systems reveals that all of them over-realize, and that conventional metrics such as CLIPScore reward the failure rather than penalize it. Prompt expansion only partly mitigates the failure. Inside the generator, the diffusion backbone attends to a mentioned target almost as much as to a requested one. The cue words that should hold the target back receive less attention than ordinary words.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.16630 [cs.CV] |
| (or arXiv:2611.16630v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2611.16630 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jiahao Huang [view email]
[v1]
Thu, 12 Nov 2026 11:03:25 UTC (2,303 KB)
来源：arXiv cs.AI · arxiv.org