OpenAIMod: OpenAI Moderation Dataset with 1,680 Examples

OpenAIMod is a dataset for content moderation, containing 1,680 text examples. It was created by RedaAlami and last updated on Hugging Face in November 2025. The dataset includes a 'prompt' column and several other columns likely representing moderation scores or categories.

Use Cases

• Training content moderation classifiers based on labeled prompts.
• Benchmarking AI safety models against the provided moderation scores.
• Analyzing patterns in flagged content based on the dataset's structure.
• Fine-tuning language models for safety alignment using the example prompts.

Strengths

• Contains 1,680 total examples for model development.
• Provides a 'prompt' column with text data for direct analysis.
• Dataset structure includes multiple columns, suggesting several moderation-related features.

Limitations

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

Provenance

Source

RedaAlami on Hugging Face.

Collection Method

Likely derived from or related to OpenAI's moderation API or systems.

Freshness

Last updated 2025-11-25 07:05:33; freshness should be verified.