diff --git a/src/gfn/gflownet/sub_trajectory_balance.py b/src/gfn/gflownet/sub_trajectory_balance.py
index 1b8da40a..f07835c3 100644
--- a/src/gfn/gflownet/sub_trajectory_balance.py
+++ b/src/gfn/gflownet/sub_trajectory_balance.py
@@ -253,31 +253,6 @@ def get_scores(
                 i,
             )
 
-            # preds = (
-            #     log_pf_trajectories_cum[i:]
-            #     - log_pf_trajectories_cum[:-i]
-            #     + current_log_state_flows
-            # )
-
-            # targets = torch.full_like(preds, fill_value=-float("inf"))
-            # assert trajectories.log_rewards is not None
-            # log_rewards = trajectories.log_rewards[trajectories.when_is_done >= i]
-
-            # targets.T[is_terminal_mask[i - 1 :].T] = log_rewards
-
-            # # For now, the targets contain the log-rewards of the ending sub trajectories
-            # # We need to add to that the log-probabilities of the backward actions up-to
-            # # the sub-trajectory's terminating state
-            # if i > 1:
-            #     targets[is_terminal_mask[i - 1 :]] += (
-            #         log_pb_trajectories_cum[i - 1 :] - log_pb_trajectories_cum[: -i + 1]
-            #     )[:-1][is_terminal_mask[i - 1 :]]
-
-            # # The following creates the targets for the non-finishing sub-trajectories
-            # targets[~full_mask[i - 1 :]] = (
-            #     log_pb_trajectories_cum[i:] - log_pb_trajectories_cum[:-i]
-            # )[:-1][~full_mask[i - 1 : -1]] + log_state_flows[i:][~sink_states_mask[i:]]
-
             flattening_mask = trajectories.when_is_done.lt(
                 torch.arange(
                     i,