Happiness Vibration BERT is a fine-tuned bert-base-uncased model designed to measure Joy Intensity (0–3) in mindfulness-inspired language.
It was developed as part of a full NLP and Generative AI pipeline exploring how linguistic tone reflects emotional well-being.

The model was trained on 1,000 programmatically generated, domain-specific comments designed to mimic the tone, language, and positive affect common in mindfulness and gratitude communities.
The dataset intentionally emphasizes higher-joy expressions to test a model’s ability to detect subtle variations within predominantly positive text.