LOKI: A Benchmark for Synthetic Data Detection Using Large Multimodal Models

A benchmark for synthetic data detection created by bczhou and released on November 5, 2024. The data supports the paper LOKI: A Comprehensive Synthetic Data Detection Benchmark using Large Multimodal Models. It is hosted on the Hugging Face platform.

Use Cases

• Benchmarking synthetic data detection algorithms based on the described multimodal benchmark.
• Training classifiers to distinguish real from synthetic content based on the benchmark's structure.
• Evaluating the performance of large multimodal models on detection tasks based on the paper's focus.

Strengths

• Data is directly linked to a named, peer-reviewed research paper (LOKI).
• The dataset was last updated on November 5, 2024, indicating recent activity.

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 and file size are unknown, which may limit suitability assessment.

Provenance

Source

bczhou on Hugging Face

Collection Method

Created for the paper 'LOKI: A Comprehensive Synthetic Data Detection Benchmark using Large Multimodal Models'.

Freshness

Last updated 2024-11-05 05:36:42.