Classify Hebrew text as neutral, positive, or negative with HeBERT’s polarity-analysis checkpoint. HeBERT Sentiment is developed by Avihay Chriqui. Adjust the available settings to shape the result for your task.
Model
avichr/heBERT_sentiment_analysis
Documentation
Use this reference to understand how every field in the playground changes the request sent to avichr/heBERT_sentiment_analysis.
Required
Type: Text
Provides one Hebrew text, a JSON list of texts, or a text-and-text-pair object for sentiment classification. The model returns neutral, positive, or negative scores based on the supplied wording; longer inputs may require tokenizer truncation.
Optional
Type: Integer
Default: 3
Range: 1–3
Sets how many sentiment labels and scores are returned, up to the model’s three labels: neutral, positive, and negative. Use 1 for only the strongest prediction or 3 to inspect the full score distribution.
Optional
Type: String
Default: None
Values: default, softmax, sigmoid, none
Chooses how raw classification scores are converted for display. Softmax produces mutually exclusive probabilities that sum to one, sigmoid scores labels independently, None leaves raw logits unchanged, and Default uses the model configuration’s expected transformation.
Optional
Type: JSON object
Default: None
Passes advanced JSON options directly to the tokenizer, such as truncation or maximum input length. Use it when the exposed controls do not cover the required preprocessing behavior, and avoid settings that remove the mask token or important source text.