fix: use textdetox/xlmr-large-toxicity-classifier as default nsfw_text model

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: rkritika1508

backend/app/core/validators/config/nsfw_text_safety_validator_config.py

@@ -10,7 +10,7 @@ class NSFWTextSafetyValidatorConfig(BaseValidatorConfig):
     threshold: float = 0.8
     validation_method: str = "sentence"
     device: Optional[str] = "cpu"
-    model_name: Optional[str] = "michellejieli/NSFW_text_classifier"
+    model_name: Optional[str] = "textdetox/xlmr-large-toxicity-classifier"
 
     def build(self):
         return NSFWText(

backend/app/tests/test_toxicity_hub_validators.py

@@ -304,7 +304,7 @@ def test_build_with_defaults(self):
         assert kwargs["threshold"] == 0.8
         assert kwargs["validation_method"] == "sentence"
         assert kwargs["device"] == "cpu"
-        assert kwargs["model_name"] == "michellejieli/NSFW_text_classifier"
+        assert kwargs["model_name"] == "textdetox/xlmr-large-toxicity-classifier"
 
     def test_build_with_custom_params(self):
         config = NSFWTextSafetyValidatorConfig(