用状态图评估大模型生成网页是否真能用，比传统方法更准。
WebRISE: Requirement-Induced State Evaluation for MLLM-Generated Web Artifacts
- 将任务需求转为可观测状态与用户意图转移图，实现无依赖执行
- 14个大模型平均仅达65.6%操作有效性和66.3%需求覆盖度
- 视频输入信号更强，且能发现文本忽略的隐式约束
现有基准通过局部证据评估多模态大模型生成的网页交互，却忽略了决定页面可用性的需求驱动状态与转换。本文提出WebRISE，将任务需求编译为交互契约图（ICGs），包含可观测状态、用户意图转移及DOM/视觉断言，支持与实现无关的浏览器执行。该框架覆盖五种输入模态（文本、Markdown、草图、图像、视频）共442个任务，包含5,495条状态转移和5,271项需求验证，区分用户显式功能与隐式产品级约束。在14个MLLM中，最强模型仅达65.6%转移有效性与66.3%需求覆盖率；视觉质量不能代表行为表现（如Qwen3.6-35B-A3B在Markdown上视觉评分80.8，但转移率仅15.5）。视频输入相比文本提升10.6个百分点的隐式覆盖，隐式约束仍普遍存续；缺陷注入实验表明，基于ICG的评分检测状态错误效率为传统检查点评估的2至16倍。
原文摘要 · Abstract (English)
Existing benchmarks for MLLM-generated web artifacts assess interaction through local evidence and miss the requirement-induced states and transitions that determine whether a page works. We introduce WebRISE, which compiles task requirements into Interaction Contract Graphs (ICGs) of observable states, user-intent transitions, and DOM/visual assertions for implementation-agnostic browser execution. WebRISE spans 442 tasks across five input modalities (Text, Markdown, Sketch, Image, Video), with 5,495 transitions and 5,271 requirement checks that separate user-stated functions from implicit product-level constraints. Across 14 MLLMs, even the strongest model reaches only 65.6% transition validity and 66.3% requirement coverage, and visual quality is no proxy for behavior (Qwen3.6-35B-A3B on Markdown: V=80.8 yet T=15.5). Video gives the strongest interaction signal (+10.6 pp implicit coverage over Text), while implicit constraints persist; defect injection shows ICG-based scoring detects state errors at 2-16x the rate of checkpoint-style evaluation.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。