arXiv cs.CL· Umar Ahmed, Maxim Dunaev, Mohammad Nour Al Awad, Sergey Ivanov, Ivan Khodnenko·· 2 天前AI 评分36
浏览器内运行 0.8B 模型检测订单静默错误:PIZZA 基准 Web 演示
Clean Tickets, Wrong Orders: Catching Silent Errors with a Language Model in Your Browser Tab
AI 导读
一个微调的 0.8B 参数模型通过 WebAssembly 完全在浏览器标签页内运行,订单文本不离开设备。在全部 1,357 条 PIZZA 测试订单上,浏览器端输出与服务器评测一致率达 99.0%,精确匹配同为 0.882。页面用移植到 TypeScript 的 PIZZA 匹配器实时打分,并对菜单外内容以“?”拒答,让访客自行寻找或制造静默错误。
正文 · 原文
Abstract:A small language model that turns a typed order into structure can fail in two ways. It can print something broken, which any system rejects. Or it can print a clean, well formed order that nobody placed. Our companion work calls the second kind a silent error, and it is the dangerous one. We present a web demo in which a fine-tuned 0.8B parameter model runs entirely inside the visitor's browser tab, through this http URL compiled to WebAssembly. No order text leaves the device. The page parses orders as the visitor types, scores every ticket with the PIZZA benchmark's own matcher ported to TypeScript, refuses anything off the menu with "?" so the model can face the open web, and lets visitors hunt for silent errors or try to cause one. Run in the browser on all 1,357 PIZZA test orders, the deployed model writes the same output as the server evaluation for 99.0% of them and reaches the same exact match (0.882); we also report its latency on 3 consumer laptops. Demo: this https URL
| Comments: | 4 pages, 4 figures, 2 tables. Demo: this https URL. Companion paper: arXiv:2611.17169 |
| Subjects: | Machine Learning (cs.LG); Computation and Language (cs.CL); Human-Computer Interaction (cs.HC) |
| ACM classes: | I.2.7; H.5.2 |
| Cite as: | arXiv:2611.18284 [cs.LG] |
| (or arXiv:2611.18284v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2611.18284 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Umar Ahmed [view email]
[v1]
Fri, 13 Nov 2026 15:13:47 UTC (22 KB)
来源:arXiv cs.CL · arxiv.org