More verbose again

src/leaderboard/build_leaderboard.py

@@ -61,6 +61,9 @@ def download_openbench():
     # download answers of different models that we trust
     download_dataset("Vikhrmodels/openbench-eval", EVAL_RESULTS_PATH)
 
+    logging.info("\nFiles openbench-eval:")
+    subprocess.run(["ls", EVAL_RESULTS_PATH], check=True)
+
     logging.info("\nInternal models in openbench-eval:")
     subprocess.run(["ls", f"{EVAL_RESULTS_PATH}/internal/"], check=True)