跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.21995 · v1

Deciding Without Pointing: A Case Study of Whether a Decision Model Localizes the Anomalies It Detects

首次提交 2026-11-17 07:27
版本更新 2026-11-17 07:27
雷达观测 2026-11-18 19:43
arXiv ↗PDF ↗

摘要

Decision models can answer a typed question about an image with a probability per option, without generating text. In our case study of Clef-flash (post-trained Qwen3.5-9B backbone) on industrial inspection benchmarks, we ask whether it localizes the anomalies it detects. We obtained uneven results with zero-shot detection: per-category image AUROC 0.924 on MVTec AD and 0.668 on MVTec AD 2. On the former, the decision head's routing attention produces 0.391 mean per-image pixel AUROC, and hiding the image tokens from it leaves detection unchanged. The multimodal LLM's backbone attention does point at the defect, however it falls below a center prior when we change question. It falls below chance if we remove textual context in the question, partly due to backbone attention leaving the object, while the head's map hardly changes. The backbone's states carry the location when probed with normal images, on the test-split images, however, the decision depends on the defect region. A larger model, Clef, that comes with a post-trained Qwen3.8-27B backbone, gives no useful map in the decision head either. We found that it's possible to recover anomaly location by asking question specifically on an image region, retraining the head or reading the backbone, but none matches dedicated few-shot localizers.

作者备注:12 pages, 5 figures, 14 tables. Preprint of the technical part of an MSc thesis (Politecnico di Milano). Code and pre-registrations are held privately until venue submission

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-17 07:27 · 最新观测版本

Deciding Without Pointing: A Case Study of Whether a Decision Model Localizes the Anomalies It Detects

Decision models can answer a typed question about an image with a probability per option, without generating text. In our case study of Clef-flash (post-trained Qwen3.5-9B backbone) on industrial inspection benchmarks, we ask whether it localizes the anomalies it detects. We obtained uneven results with zero-shot detection: per-category image AUROC 0.924 on MVTec AD and 0.668 on MVTec AD 2. On the former, the decision head's routing attention produces 0.391 mean per-image pixel AUROC, and hiding the image tokens from it leaves detection unchanged. The multimodal LLM's backbone attention does point at the defect, however it falls below a center prior when we change question. It falls below chance if we remove textual context in the question, partly due to backbone attention leaving the object, while the head's map hardly changes. The backbone's states carry the location when probed with normal images, on the test-split images, however, the decision depends on the defect region. A larger model, Clef, that comes with a post-trained Qwen3.8-27B backbone, gives no useful map in the decision head either. We found that it's possible to recover anomaly location by asking question specifically on an image region, retraining the head or reading the backbone, but none matches dedicated few-shot localizers.

查看此版本解读 →查看此版本原文 ↗