README

README.md

@@ -9,7 +9,7 @@ base_model:
 
 ## Overview
 
-Nous Research presents Minos, a lightweight classifier designed to detect refusals in text. Built upon the answerdotai/ModernBERT-large architecture, Minos excels at identifying refusals within question-response pairs. We leverage Minos internally to ensure our synthetic responses are free from refusals. Trained on over 350,000 single and multi-turn question-response pairs, we are now releasing the Minos classifier to the community. We hope it proves valuable for identifying and managing refusals in your own applications!
+Nous Research presents Minos, a lightweight classifier designed to detect refusals in text. Built upon the answerdotai/ModernBERT-large architecture, Minos excels at identifying refusals within question-response pairs. We leverage Minos internally to ensure our synthetic responses are free from refusals and hope it proves valuable for identifying and managing refusals in your applications!
 
 ## Model Architecture
 

examples/inference_server.py

@@ -59,7 +59,7 @@ class MinosRefusalClassifier:
         """Single turn prediction for simple use cases"""
         return self.predict_multi_turn([{"user": user_text, "assistant": assistant_text}])
 
-# Usage example
+
 if __name__ == "__main__":
     # Load from Hugging Face Hub
     classifier = MinosRefusalClassifier()