patrickjamesmarcellana/synthetic-filipino-sarcasm-detection

Synthetic and Limited Real-World Filipino Sarcasm Detection Dataset This dataset is composed of Filipino sarcastic and non-sarcastic tweets, divided into two categories: LLM-generated (synthetic) data and real-world data. Synthetic Data Information Two large language models were used to generate sarcastic and non-sarcastic tweets for the dataset: GPT-4o and Gemini 2.0 Flash. The dataset is composed of 504 sarcastic tweets (252 for each LLM) and 504 non-sarcastic… See the full description on the dataset page: https://huggingface.co/datasets/patrickjamesmarcellana/synthetic-filipino-sarcasm-detection.