added help text to user inputs

app.py

@@ -130,6 +130,7 @@ st.markdown("# 🤗 Leaderboards")
 
 only_verified_results = st.sidebar.checkbox(
     "Filter for Verified Results",
+    help="Select this checkbox if you want to see only results produced by the Hugging Face model evaluator, and no self-reported results."
 )
 
 selectable_tasks = list(set(dataframe.pipeline_tag))
@@ -140,6 +141,7 @@ selectable_tasks.sort(key=lambda name: name.lower())
 task = st.sidebar.selectbox(
     "Task",
     ["-any-"] + selectable_tasks,
+    help="Filter the selectable datasets by task. Leave as \"-any-\" to see all selectable datasets."
 )
 
 if task != "-any-":
@@ -163,6 +165,7 @@ dataset = st.sidebar.selectbox(
     "Dataset",
     selectable_datasets,
     index=selectable_datasets.index(default_dataset) if default_dataset in selectable_datasets else 0,
+    help="Select a dataset to see the leaderboard!"
 )
 
 dataframe = dataframe[dataframe.only_verified == only_verified_results]
@@ -178,6 +181,7 @@ if len(dataset_df) > 0:
     config = st.sidebar.selectbox(
         "Config",
         selectable_configs,
+        help="Filter the results on the current leaderboard by the dataset config. Self-reported results might not report the config, which is why \"-unspecified-\" is an option."
     )
     dataset_df = dataset_df[dataset_df.config == config]
 
@@ -185,6 +189,7 @@ if len(dataset_df) > 0:
     split = st.sidebar.selectbox(
         "Split",
         selectable_splits,
+        help="Filter the results on the current leaderboard by the dataset split. Self-reported results might not report the split, which is why \"-unspecified-\" is an option."
     )
     dataset_df = dataset_df[dataset_df.split == split]
 
@@ -196,6 +201,7 @@ if len(dataset_df) > 0:
     sorting_metric = st.sidebar.radio(
         "Sorting Metric",
         selectable_metrics,
+        help="Select the metric to sort the leaderboard by. Click on the metric name in the leaderboard to reverse the sorting order."
     )
 
     st.markdown(