WebRISE

WebRISE is a benchmark for requirement-induced state evaluation of MLLM-generated web artifacts. The release contains task specifications, multimodal inputs, ground-truth HTML artifacts, and Interaction Contract Graphs (ICGs) for browser-based evaluation. ## Contents - `requirements_full.json`: full task catalog and requirement annotations. - `data/<task_id>/` on Hugging Face, or `<task_id>/` in this local release: one folder per task. - `<task_id>/icg.json`: Interaction Contract Graph with states, transitions, DOM assertions, visual postconditions, and requirement coverage mappings. - `<task_id>/*.html`: ground-truth executable HTML artifact. - `<task_id>/Text`, `<task_id>/MD`, `<task_id>/Sketch`, `<task_id>/Image`, `<task_id>/Video`: modality-specific input specifications when available. The release includes 442 tasks across five input modalities. ## Citation If you use WebRISE in your research, please cite our paper: ```bibtex @misc{meng2026webriserequirementinducedstateevaluation, title={WebRISE: Requirement-Induced State Evaluation for MLLM-Generated Web Artifacts}, author={Yuxin Meng and Yuhan Suo and Junjie Wang and Yuhan Sun and Yiyao Yu and Ruixu Zhang and Ruining Hu and Yubin Wang and Shouwei Ruan and Bin Wang and Yuxiang Zhang and Yujiu Yang}, year={2026}, eprint={2606.03220}, archivePrefix={arXiv}, primaryClass={cs.CL}, url={https://arxiv.org/abs/2606.03220}, } ```