LADBench: A Benchmark for Logical Anomaly Detection in Images

LAD-Bench is a benchmark of more than 1,000 curated synthetic images designed to test the logical reasoning capabilities of Vision Language Models. It was created by SahasraK and introduced to address gaps in evaluating physical and social common sense for open-world AI deployment. The dataset was last updated on June 16, 2026.

Use Cases

• Benchmarking logical reasoning in VLMs based on synthetic images with anomalies.
• Training models to detect physical and social common sense violations in images.
• Evaluating autonomous AI systems for open-world deployment scenarios.

Strengths

• Contains more than 1,000 curated synthetic images.
• Specifically designed to test logical reasoning, not just visual errors.

Limitations

• Dataset size and file formats are unknown.
• Column-level documentation is absent; field semantics must be inferred after download.
• Data may reflect bias inherent to its synthetic generation method.

Provenance

Source

huggingface

Collection Method

Synthetic image curation.

Freshness

Last updated 2026-06-16 04:23:11; freshness should be verified.