Limitations

Scores are a classifier's reading of the text, not a measurement of what anyone feels.

• A score is the classifier's confidence that a label applies to the text. It is not an intensity.
• The model was trained on Reddit comments, so it may read other kinds of text, such as AI replies, less reliably.
• Every result names the model and version that produced it, because a different version can read the same text differently.

Known weaknesses

• Rare emotions are read less reliably than common ones. Grief, pride, relief, nervousness and embarrassment have few examples in the training data.
• Sarcasm, quotation and very short text are easy to misread.
• The models read English only, and were trained on Reddit comments, so they may read other kinds of text, such as AI replies, less reliably.

Each model's card on Hugging Face reports its test results for every label. See Models.

Uses to avoid

• Decisions about a person: hiring, grading, discipline, credit, insurance, policing, immigration, or any medical or mental-health assessment.
• Reading a person: a score is not evidence of how a particular person feels.
• Watching people: monitoring employees, students or anyone who hasn't agreed to it.
• Proof: a score is one model's reading. It doesn't settle what a text means, and it says nothing about whether an AI model has inner experience.