# Social media content moderation

A Portuguese hate speech detection model fine-tuned based on neuralmind/bert-large-portuguese-cased, achieving an accuracy of 98.92%

This model is fine-tuned based on `dbmdz/bert-base-turkish-cased` and is used for binary toxicity classification of Turkish texts.

A toxicity detection model fine-tuned based on BETO (Spanish BERT base model), designed to identify toxic content in Spanish texts.

This model is a BERT-architecture optimized implicit hate speech detector, specifically trained and fine-tuned for English implicit hate speech.

IndoBERTweet-HateSpeech is a model fine-tuned on the IndoToxic2024 dataset for identifying hate speech in Indonesian, with high accuracy and F1 score.

MetaHateBERT is a text classification model based on the BERT architecture, specifically designed for detecting hate speech.

A BERT-based Turkish text classification model for detecting offensive language in text

A hate/offensive speech detector based on the lightweight DistilBERT model, achieving efficient recognition through massive text pre-training and fine-tuning with proprietary datasets

A hate speech classifier fine-tuned on the BERT-large Portuguese model, achieving 78.14% accuracy on the hate_speech_portuguese dataset

A Turkish hate speech detection model fine-tuned based on BERTurk, capable of identifying neutral, offensive, and hate speech texts.

This model is a fine-tuned version based on ibert-roberta-base, specifically designed for detecting abusive or threatening speech.

Ultra-high-precision multi-class image classifier for accurate sensitive content detection

A Korean text classification model based on ELECTRA architecture, designed to determine if comments and chat content contain sensitive information

This is a Turkish offensive language detection model based on the BERT architecture, fine-tuned with Twitter data.

This is a model fine-tuned on a tweet hate speech detection dataset, designed to identify and classify hate speech on social media.

This is a monolingual model for hate speech classification in Italian social media content, built upon the ALBERTO pre-trained language model.

A toxicity detection model for Spanish political tweets fine-tuned based on xlm-roberta-base, capable of identifying toxic and highly toxic content in tweets.

A monolingual model for classifying hate speech in Dutch social media content, based on the BERTje pre-trained model

This model is used to detect hate speech in Polish, fine-tuned based on a multilingual BERT model and trained using only English data.

This model is designed to detect hate speech in Indonesian, fine-tuned based on multilingual BERT and trained exclusively with Indonesian data.

A BERT-based text classification model for detecting hate speech and offensive content, with rationale prediction capability

This model is used to detect offensive content in Kannada mixed code, fine-tuned based on the XLM-Roberta-Base pre-trained model, and performed excellently in the EACL 2021 shared task.

A Danish offensive language detection model based on the ELECTRA architecture, used to determine if text is offensive.

HateXplain is a text classification model for detecting hate speech, offensive content, and normal content, trained on data from Gab and Twitter, with enhanced performance through human-annotated rationale.

A hate speech detection model based on Danish BERT, used to identify offensive content in Danish text.