Research map / Video / Video-language understanding
IWR-Bench: Can LVLMs reconstruct interactive webpage from a user interaction video?
Yang Chen, Minghao Liu, Yufan Shen, Yunwen Li, Tianyuan Huang, Xinyu Fang, Tianyu Zheng, Wenxuan Huang, Cheng Yang, Licheng Wen, Xuemeng Yang, Daocheng Fu and 8 more
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| enable thinking mode on Claude-Sonnet-4 | Claude-Sonnet-4 · IWR-Bench | Final Score: 34.00 -> 34.62 | Improved |
| enable thinking mode on Claude-Sonnet-4 | Claude-Sonnet-4 · IWR-Bench | Interactive Functionality Score: 22.29 -> 23.65 | Improved |
| enable thinking mode on Claude-Opus-4 | Claude-Opus-4 · IWR-Bench | Final Score: 33.33 -> 34.13 | Improved |
| enable thinking mode on Claude-Opus-4 | Claude-Opus-4 · IWR-Bench | Interactive Functionality Score: 21.83 -> 23.61 | Improved |
| enable thinking mode on Gemini-2.5-Pro | Gemini-2.5-Pro · IWR-Bench | Final Score: 30.31 -> 30.36 | No clear effect |
| enable thinking mode on Gemini-2.5-Pro | Gemini-2.5-Pro · IWR-Bench | Interactive Functionality Score: 20.51 -> 21.65 | Improved |
| enable thinking mode on Gemini-2.5-Pro | Gemini-2.5-Pro · IWR-Bench | Visual Fidelity Score: 53.18 -> 50.67 | Hurt |
| move from static content (L1) to interactive tasks (L2 simple state manipulation) | GPT-5 · IWR-Bench | Final Score: 61.85 -> 35.43 | Hurt |
| scale up from Qwen2.5-VL-7B to 72B | Qwen2.5-VL · IWR-Bench | Final Score: 15.47 -> 23.61 | Improved |
| use video-specialized model instead of general LVLM (VideoLLaMA3-7B vs Qwen2.5-VL-7B) | 7-8B open-source models · IWR-Bench | Final Score: 13.67 vs 15.47 | No clear effect |
| set final-score weighting alpha to 0.7 selected by human-alignment grid search (ablation) | Final Score metric · IWR-Bench human alignment | Spearman correlation: rho=0.65 | Improved |
Closest papers by meaning
- Animation2Code: Evaluating Temporal Visual Reasoning in Video-to-Code Generation (NeurIPS 2026 · similarity 0.89)
- Vision2Web: A Hierarchical Benchmark for Visual Website Development with Agent Verification (ICML 2026 · similarity 0.88)
- Web2Code: A Large-scale Webpage-to-Code Dataset and Evaluation Framework for Multimodal LLMs (NeurIPS 2024 · similarity 0.88)
- VisCodex: Unified Multimodal Code Generation via Merging Vision and Coding Models (ICLR 2026 · similarity 0.86)
- VideoWebArena: Evaluating Long Context Multimodal Agents with Video Understanding Web Tasks (ICLR 2025 · similarity 0.86)
- Harnessing Webpage UIs for Text-Rich Visual Understanding (ICLR 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
