Target-Group Classifier
The BERT-based counter-argument classifier is finetuned on the combined dataset of CONAN and CrowdCounter for classifying whether a sequence is about one or multiple of the 8 target group, based on the sexism detector ThinkCERCA/counterargument_hugging
It classifies either a given sentence is a valid counter-narrative.
Uses
The model is intended for classifying LM-generated dialogue responses, and evaluating their validity as counter-narrative.
- Downloads last month
- 14
Model tree for tum-nlp/counter-narrative-classifier
Base model
ThinkCERCA/counterargument_hugging