bert-base-uncased-hatexplain
| Property | Value |
|---|---|
| License | Apache 2.0 |
| Paper | HateXplain Paper |
| Downloads | 93,896 |
| Framework | PyTorch, JAX |
What is bert-base-uncased-hatexplain?
bert-base-uncased-hatexplain is a specialized BERT-based model designed for detecting and classifying hate speech in social media content. Built on the BERT architecture, this model has been specifically trained to categorize text into three distinct classes: Hatespeech, Offensive, or Normal. What makes it unique is its training approach that incorporates human rationales alongside data from Twitter and Gab platforms.
Implementation Details
The model builds upon the BERT base uncased architecture and has been fine-tuned using the HateXplain dataset. The training process incorporates human annotations and rationales to improve classification accuracy and provide more contextual understanding of hate speech patterns.
- Leverages transformer architecture for contextual understanding
- Incorporates human rationales in training data
- Supports both PyTorch and JAX frameworks
- Trained on diverse social media content from Twitter and Gab
Core Capabilities
- Three-way classification: Hatespeech, Offensive, and Normal content
- Context-aware text analysis
- Explainable classifications through human rationales
- Efficient processing of social media content
Frequently Asked Questions
Q: What makes this model unique?
This model stands out due to its incorporation of human rationales during training, which helps in providing more accurate and explainable hate speech detection. It's also been trained on a diverse dataset combining content from both Twitter and Gab platforms.
Q: What are the recommended use cases?
The model is ideal for content moderation systems, social media platforms, and research applications requiring automated detection of hate speech and offensive content. It's particularly useful when explainability of classifications is important.
