diff --git a/pearl/policy_learners/contextual_bandits/neural_linear_bandit.py b/pearl/policy_learners/contextual_bandits/neural_linear_bandit.py
index fc4e3e24..b47e941d 100644
--- a/pearl/policy_learners/contextual_bandits/neural_linear_bandit.py
+++ b/pearl/policy_learners/contextual_bandits/neural_linear_bandit.py
@@ -63,7 +63,7 @@ def __init__(
         learning_rate: float = 0.0003,
         l2_reg_lambda_linear: float = 1.0,
         state_features_only: bool = False,
-        loss_type: str = "mse",  # one of the LOSS_TYPES names, e.g., mse, mae, xentropy
+        loss_type: str = "mse",  # one of the LOSS_TYPES names: [mse, mae, cross_entropy]
         output_activation_name: str = "linear",
         use_batch_norm: bool = False,
         use_layer_norm: bool = False,