MrRamyg/orca-moderation-corpus

Orca Moderation Corpus

The Orca Moderation Corpus is a public dataset designed for text content moderation tasks. It contains examples of profanity, hate speech, violent language, and offensive content across multiple contexts. This dataset can be used to train, evaluate, and benchmark models for detecting toxic and harmful text.

⚠️ Warning: This dataset contains explicit, offensive, and potentially disturbing content. Use responsibly and in compliance with ethical AI guidelines.

Dataset Summary

• Name: Orca Moderation Corpus
• Version: 1.0.0
• License: CC BY 4.0
• Language: Primarily English
• Size: Varies (depending on your upload)
• Intended Use: Research, model training, moderation evaluation

Dataset Structure

Each row in the dataset is a text sample labeled for content type. Example schema:

Column	Description
text	The raw text containing profanity, hate, or violence
profanity	Boolean indicating presence of profanity
hate	Boolean indicating presence of hate speech
violence	Boolean indicating presence of violent content
nsfw	Boolean indicating sexually explicit/offensive content (optional)

Example Data

{
  "text": "F*ck you, you idiot!",
  "profanity": true,
  "hate": false,
  "violence": false,
  "nsfw": false
}

{
  "text": "I will make you pay for this!",
  "profanity": false,
  "hate": false,
  "violence": true,
  "nsfw": false
}

Usage

from datasets import load_dataset

dataset = load_dataset("MrRamyg/orca-moderation-corpus")

# Example: filter all violent texts
violent_texts = dataset["train"].filter(lambda x: x["violence"])
print(violent_texts[:5])

Notes

• The dataset includes explicit language and offensive content to reflect real-world moderation challenges.
• Recommended for developers, researchers, and model trainers working on NLP moderation systems.
• Handle data responsibly and respect ethical AI standards.

Citation

@misc{orca_moderation_corpus_2026,
  title={Orca Moderation Corpus},
  author={MrRamyg},
  year={2026},
  howpublished={\url{https://huggingface.co/datasets/MrRamyg/orca-moderation-corpus}}
}