Dataset · google
Wikipedia comments with toxicity labels
google/jigsaw_toxicity_pred
A classic dataset by Google for classifying toxic behavior in English texts.
What's inside
Contains English Wikipedia comment texts labeled by humans for toxicity. The dataset size ranges from 100 thousand to 1 million records.
What you can do with it
- Training classifiers for automatic comment moderation
- Configuring safety guards for language models
- Tasks
- Text classification
- Languages
- English
- Size
- 100K–1M rows
- Downloads this month
- 1.1K
- Likes
- 35
Models trained on it
- s-nlp/roberta_toxicity_classifierText classification 60.2K
- garak-llm/roberta_toxicity_classifier 18.2K
- AssistantsLab/Tiny-Toxic-DetectorText classification 59
- s-nlp/roberta_first_toxicity_classifierText classification 39
- StrictlyInsecure/discord-moderation-minilmText classification 39
- StrictlyInsecure/discord-moderation-minilm-l12Text classification 22
- enguard/tiny-guard-4m-en-prompt-toxicity-binary-jigsawText classification 20
- enguard/medium-guard-128m-xx-prompt-toxicity-binary-jigsawText classification 17
How to load it
from datasets import load_dataset
ds = load_dataset("google/jigsaw_toxicity_pred")