update

src/backend/envs.py

@@ -22,7 +22,8 @@ class Tasks(Enum):
     # task1 = Task("logiqa", "acc_norm", "LogiQA")
     task0 = Task("nq_open", "em", "NQ Open", 64)  # 64, as in the ATLAS paper
     task1 = Task("triviaqa", "em", "TriviaQA", 64)  # 64, as in the ATLAS paper
-    task2 = Task("truthfulqa_mc", "mc2", "TruthfulQA", 0)  # TruthfulQA is intended as a zero-shot benchmark [5, 47]. https://owainevans.github.io/pdfs/truthfulQA_lin_evans.pdf
+    task2 = Task("truthfulqa_mc1", "mc1", "TruthfulQA MC1", 0)
+    task3 = Task("truthfulqa_mc2", "mc2", "TruthfulQA MC2", 0)  # TruthfulQA is intended as a zero-shot benchmark [5, 47]. https://owainevans.github.io/pdfs/truthfulQA_lin_evans.pdf
 
 
 # NUM_FEWSHOT = 64  # Change with your few shot