Fix: Correct attribute references for error detection/correction metrics

- Changed error_detection_count to errors_detected (actual attribute name)
- Changed error_correction_count to errors_corrected (actual attribute name)
- Fixed both app.py (main) and RGBMetrics/app.py (mirror)
- Resolves 'EvaluationResult' object has no attribute 'error_detection_count' error

app.py

@@ -174,9 +174,9 @@ def run_evaluation_background(selected_models, selected_tasks, max_samples, api_
                     "accuracy": r.accuracy,
                     "rejected": r.rejected,
                     "rejection_rate": r.rejection_rate,
-                    "error_detection_count": r.error_detection_count,
+                    "error_detection_count": r.errors_detected,
                     "error_detection_rate": r.error_detection_rate,
-                    "error_correction_count": r.error_correction_count,
+                    "error_correction_count": r.errors_corrected,
                     "error_correction_rate": r.error_correction_rate,
                 }
                 for r in results