diff --git a/nbs/src/nixtla_client.ipynb b/nbs/src/nixtla_client.ipynb
index 5599e663..94460f7d 100644
--- a/nbs/src/nixtla_client.ipynb
+++ b/nbs/src/nixtla_client.ipynb
@@ -1161,7 +1161,6 @@
     "        )\n",
     "        out = ufp.assign_columns(out, 'TimeGPT', resp['mean'])\n",
     "        out = _maybe_add_intervals(out, resp['intervals'])\n",
-    "        out = _maybe_convert_level_to_quantiles(out, quantiles)\n",
     "        if add_history:\n",
     "            in_sample_df = _parse_in_sample_output(\n",
     "                in_sample_output=in_sample_resp,\n",
@@ -1173,6 +1172,7 @@
     "            )\n",
     "            in_sample_df = ufp.drop_columns(in_sample_df, target_col)\n",
     "            out = ufp.vertical_concat([in_sample_df, out])\n",
+    "        out = _maybe_convert_level_to_quantiles(out, quantiles)\n",
     "        self._maybe_assign_feature_contributions(\n",
     "            expected_contributions=feature_contributions,\n",
     "            resp=resp,\n",
@@ -1184,8 +1184,10 @@
     "            sort_idxs = ufp.maybe_compute_sort_indices(out, id_col=id_col, time_col=time_col)\n",
     "            if sort_idxs is not None:\n",
     "                out = ufp.take_rows(out, sort_idxs)\n",
+    "                out = ufp.drop_index_if_pandas(out)\n",
     "                if hasattr(self, 'feature_contributions'):\n",
     "                    self.feature_contributions = ufp.take_rows(self.feature_contributions, sort_idxs)\n",
+    "                    self.feature_contributions = ufp.drop_index_if_pandas(self.feature_contributions)\n",
     "        out = _maybe_drop_id(df=out, id_col=id_col, drop=drop_id)\n",
     "        self._maybe_assign_weights(weights=resp['weights_x'], df=df, x_cols=x_cols)\n",
     "        return out\n",
@@ -1833,8 +1835,10 @@
     "        **kwargs\n",
     "    )\n",
     "    assert all(col in df_qls.columns for col in exp_q_cols)\n",
+    "    assert not any('-lo-' in col for col in df_qls.columns)\n",
     "    # test monotonicity of quantiles\n",
-    "    df_qls.apply(lambda x: x.is_monotonic_increasing, axis=1).sum() == len(exp_q_cols)\n",
+    "    for c1, c2 in zip(exp_q_cols[:-1], exp_q_cols[1:]):\n",
+    "        assert df_qls[c1].lt(df_qls[c2]).all()\n",
     "test_method_qls(nixtla_client.forecast)\n",
     "test_method_qls(nixtla_client.forecast, add_history=True)\n",
     "test_method_qls(nixtla_client.cross_validation)"
@@ -2394,8 +2398,8 @@
     "anom_inferred_df_index = nixtla_client.detect_anomalies(df_ds_index)\n",
     "fcst_inferred_df = nixtla_client.forecast(df_[['ds', 'unique_id', 'y']], h=10)\n",
     "anom_inferred_df = nixtla_client.detect_anomalies(df_[['ds', 'unique_id', 'y']])\n",
-    "pd.testing.assert_frame_equal(fcst_inferred_df_index, fcst_inferred_df, atol=1e-3)\n",
-    "pd.testing.assert_frame_equal(anom_inferred_df_index, anom_inferred_df, atol=1e-3)\n",
+    "pd.testing.assert_frame_equal(fcst_inferred_df_index, fcst_inferred_df, atol=1e-4, rtol=1e-3)\n",
+    "pd.testing.assert_frame_equal(anom_inferred_df_index, anom_inferred_df, atol=1e-4, rtol=1e-3)\n",
     "df_ds_index = df_ds_index.groupby('unique_id').tail(80)\n",
     "for freq in ['Y', 'W-MON', 'Q-DEC', 'H']:\n",
     "    df_ds_index.index = np.concatenate(\n",
@@ -2405,7 +2409,7 @@
     "    fcst_inferred_df_index = nixtla_client.forecast(df_ds_index, h=10)\n",
     "    df_test = df_ds_index.reset_index()\n",
     "    fcst_inferred_df = nixtla_client.forecast(df_test, h=10)\n",
-    "    pd.testing.assert_frame_equal(fcst_inferred_df_index, fcst_inferred_df, atol=1e-3)"
+    "    pd.testing.assert_frame_equal(fcst_inferred_df_index, fcst_inferred_df, atol=1e-4, rtol=1e-3)"
    ]
   },
   {
@@ -2547,7 +2551,9 @@
     "\n",
     "pd.testing.assert_frame_equal(\n",
     "    timegpt_anomalies_df_1,\n",
-    "    timegpt_anomalies_df_2 \n",
+    "    timegpt_anomalies_df_2,\n",
+    "    atol=1e-4,\n",
+    "    rtol=1e-3,\n",
     ")"
    ]
   },
diff --git a/nixtla/nixtla_client.py b/nixtla/nixtla_client.py
index 2d11c4da..2298d21f 100644
--- a/nixtla/nixtla_client.py
+++ b/nixtla/nixtla_client.py
@@ -1093,7 +1093,6 @@ def forecast(
         )
         out = ufp.assign_columns(out, "TimeGPT", resp["mean"])
         out = _maybe_add_intervals(out, resp["intervals"])
-        out = _maybe_convert_level_to_quantiles(out, quantiles)
         if add_history:
             in_sample_df = _parse_in_sample_output(
                 in_sample_output=in_sample_resp,
@@ -1105,6 +1104,7 @@ def forecast(
             )
             in_sample_df = ufp.drop_columns(in_sample_df, target_col)
             out = ufp.vertical_concat([in_sample_df, out])
+        out = _maybe_convert_level_to_quantiles(out, quantiles)
         self._maybe_assign_feature_contributions(
             expected_contributions=feature_contributions,
             resp=resp,
@@ -1118,10 +1118,14 @@ def forecast(
             )
             if sort_idxs is not None:
                 out = ufp.take_rows(out, sort_idxs)
+                out = ufp.drop_index_if_pandas(out)
                 if hasattr(self, "feature_contributions"):
                     self.feature_contributions = ufp.take_rows(
                         self.feature_contributions, sort_idxs
                     )
+                    self.feature_contributions = ufp.drop_index_if_pandas(
+                        self.feature_contributions
+                    )
         out = _maybe_drop_id(df=out, id_col=id_col, drop=drop_id)
         self._maybe_assign_weights(weights=resp["weights_x"], df=df, x_cols=x_cols)
         return out