arXiv cs.CL· Jiacheng Ding, Yanliang Qi, Cong Guo, Peihua Mai·· 11 小时前AI 评分38
See, Think, Use：诊断视觉语言模型的感知与推理失败
See, Think, Use: Diagnosing Perception and Reasoning Failures in VLMs
AI 导读
研究者提出 See, Think, Use 协议，将视觉语言模型的每个答案归因到感知、推理或对感知的利用三个阶段，无需访问模型权重，每道题仅需额外提问两次。在两类模型家族、六个数据集上，该协议显示有效调节手段随任务类型而变；答案通常遵循模型自身感知，但在图像与常见先验相矛盾的样本上会失效，先定位再作答可在定位框正确时修复该失败。按测得阶段分配测试时算力，成本低于最佳统一策略且接近其效果。
正文 · 原文
Abstract:We present See, Think, Use, a protocol that traces each answer of a vision-language model to perception, reasoning or the use of perception. The diagnosis shows a developer which stage to improve and directs test-time compute to the stage that fails. Benchmarks score the final answer, and an answer that rests on the image earns the same credit as a guess or a memorized prior. The protocol asks the model two extra questions per item and needs no access to its weights. A localization probe asks for a box around every instance and measures seeing, and a text-replacement probe answers from a description and measures thinking. Use is the rate at which the answer follows the model's own perception across counterfactual image pairs. The rate needs no answer labels, because the question and its prior are the same on both images. On two model families and six datasets the protocol shows that the knob that helps follows the task family. Answers follow the model's perception except on images that contradict a common prior, and asking the model to localize first repairs this failure when its boxes are right. Allocating compute by the measured stages comes close to the best uniform policy at lower cost and beats a random allocation of the same compute wherever a knob raises the short stage. We will release the code, the prompts and the per-question outputs.
| Comments: | 10 pages, 4 figures, 8 tables |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2611.01467 [cs.CV] |
| (or arXiv:2611.01467v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2611.01467 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Cong Guo [view email]
[v1]
Sun, 1 Nov 2026 02:56:06 UTC (2,943 KB)
来源：arXiv cs.CL · arxiv.org