LLM Text Quality & Toxicity Checker

A simple web application for checking the quality and toxicity of English text.
The tool uses BERT-based models (unitary/toxic-bert and dehatebert-mono-english), together with a custom keyword list with toxicity weights.

Features

• Grammar, spelling, and readability scoring
• Toxicity detection using both transformer models and keyword filtering
• Easy-to-use Streamlit web interface
• Fast scoring and export for dataset curation or content moderation

Quick Start

1. Clone the repo:

   git clone https://github.com/kullanici_adin/llm-text-quality-toxicity.git
   

2. Install requirements:

   pip install -r requirements.txt
   

3. Run the app:

   streamlit run app.py
   

Models Used

• unitary/toxic-bert
• Hate-speech-CNERG/dehatebert-mono-english

---

This project helps you filter out harmful or low-quality English text quickly and easily.