clean up repo

src/populate.py

@@ -6,6 +6,7 @@ import yaml
 from sklearn.metrics import cohen_kappa_score
 import numpy as np
 from datasets import load_dataset
+from envs import TOKEN
 
 TYPES = ["str", "number", "number", "number", "number", "number"]
 
@@ -84,7 +85,7 @@ def load_leaderboard() -> pd.DataFrame:
     """Loads the leaderboard from the file system"""
     with open("./data/models.yaml") as fp:
         models = yaml.safe_load(fp)
-    human_responses = load_dataset("salesforce/instrusum", "human_eval_pairwise")["data"]
+    human_responses = load_dataset("salesforce/instrusum", "human_eval_pairwise", token=TOKEN)["data"]
     human_responses = [x for x in human_responses]
 
     predictions = {k: [] for k in ["Model", "Accuracy", "Agreement", "Self-Accuracy", "Self-Agreement"]}