Update app.py

app.py

@@ -247,7 +247,7 @@ def load_dataset_samples():
                     "response_b": responses[model_b],
                     "model_a": model_a,
                     "model_b": model_b,
-                    "dataset": item.get("dataset", "unknown")
+                    "dataset": "NTNU"
                 })
 
         extra_dataset = load_dataset("ltg/fluency-generations", split="test", token=HF_TOKEN)
@@ -256,8 +256,8 @@ def load_dataset_samples():
             sample_id = item["sample_id"]
             prompt = item["prompt"]
             responses = item["responses"]
-            model_a, model_b = MODEL_PAIRS[i]
-            model_a, model_b = model_a, model_b if i % 2 == 0 else model_b, model_a
+            model_a, model_b = MODEL_PAIRS[i % len(MODEL_PAIRS)]
+            model_a, model_b = (model_a, model_b) if i % 2 == 0 else (model_b, model_a)
             extra_pairwise_samples.append({
                 "id": f"{sample_id}_{model_a}_vs_{model_b}",
                 "original_id": sample_id,
@@ -266,7 +266,7 @@ def load_dataset_samples():
                 "response_b": responses[model_b],
                 "model_a": model_a,
                 "model_b": model_b,
-                "dataset": item.get("dataset", "unknown")
+                "dataset": "training_examples"
             })
         
         return pairwise_samples, extra_pairwise_samples