Update app.py

app.py

@@ -164,7 +164,7 @@ SUMMARISATION = Task(name="summarisation", metric="bertscore")
 KNOWLEDGE = Task(name="knowledge", metric="mcc")
 REASONING = Task(name="reasoning", metric="mcc")
 GRAMMAR = Task(name="grammar", metric="mcc")
-QUESTION_ANSWERING = Task(name="question answering", metric="em")
+READING_COMPREHENSION = Task(name="reading comprehension", metric="em")
 TEXT_CLASSIFICATION = Task(name="text classification", metric="mcc")
 INFORMATION_EXTRACTION = Task(name="information extraction", metric="micro_f1_no_misc")
 ALL_TASKS = [obj for obj in globals().values() if isinstance(obj, Task)]
@@ -203,13 +203,13 @@ DATASETS = [
     Dataset(name="scala-de", language=GERMAN, task=GRAMMAR),
     Dataset(name="scala-nl", language=DUTCH, task=GRAMMAR),
     Dataset(name="scala-en", language=ENGLISH, task=GRAMMAR),
-    Dataset(name="scandiqa-da", language=DANISH, task=QUESTION_ANSWERING),
-    Dataset(name="norquad", language=NORWEGIAN, task=QUESTION_ANSWERING),
-    Dataset(name="scandiqa-sv", language=SWEDISH, task=QUESTION_ANSWERING),
-    Dataset(name="nqii", language=ICELANDIC, task=QUESTION_ANSWERING),
-    Dataset(name="germanquad", language=GERMAN, task=QUESTION_ANSWERING),
-    Dataset(name="squad", language=ENGLISH, task=QUESTION_ANSWERING),
-    Dataset(name="squad-nl", language=DUTCH, task=QUESTION_ANSWERING),
+    Dataset(name="scandiqa-da", language=DANISH, task=READING_COMPREHENSION),
+    Dataset(name="norquad", language=NORWEGIAN, task=READING_COMPREHENSION),
+    Dataset(name="scandiqa-sv", language=SWEDISH, task=READING_COMPREHENSION),
+    Dataset(name="nqii", language=ICELANDIC, task=READING_COMPREHENSION),
+    Dataset(name="germanquad", language=GERMAN, task=READING_COMPREHENSION),
+    Dataset(name="squad", language=ENGLISH, task=READING_COMPREHENSION),
+    Dataset(name="squad-nl", language=DUTCH, task=READING_COMPREHENSION),
     Dataset(name="nordjylland-news", language=DANISH, task=SUMMARISATION),
     Dataset(name="mlsum", language=GERMAN, task=SUMMARISATION),
     Dataset(name="rrn", language=ICELANDIC, task=SUMMARISATION),
@@ -671,11 +671,6 @@ def fetch_results() -> dict[Language, pd.DataFrame]:
         for record in records:
             model_name = record["model"]
 
-            # Manual fix for OpenAI models: Only keep the validation split results
-            if "gpt-3.5" in model_name or "gpt-4" in model_name:
-                if not record.get("validation_split", False):
-                    continue
-
             dataset_name = record["dataset"]
             if dataset_name in possible_dataset_names:
                 dataset = next(