RUC: Real UI Clickboxes for Deceptive Interface Understanding

RUC: Real UI Clickboxes is a benchmark dataset for evaluating vision-language models and web agents on their ability to understand and resist deceptive user interfaces. The dataset accompanies an ACL submission and is released by the DUDE-Framework. It was last updated on Hugging Face on May 13, 2026.

Use Cases

• Benchmarking model performance on deceptive UI detection based on the described focus.
• Training web agents to identify misleading clickable elements based on the dataset's purpose.
• Researching human-computer interaction patterns in adversarial web environments based on the paper's topic.

Strengths

• Dataset is associated with a peer-reviewed ACL submission, suggesting a research foundation.
• Explicitly designed for the specific task of evaluating deceptive UI understanding for web agents.
• Last update timestamp (2026-05-13 13:40:17) is provided.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count, file formats, and license information are unknown, which may limit suitability assessment.

Provenance

Source

DUDE-Framework

Collection Method

Likely collected or synthesized for research on deceptive web interfaces, as described in the associated paper.

Freshness

Last updated 2026-05-13 13:40:17; freshness should be verified.