diff --git a/torchy_baselines/ppo/ppo.py b/torchy_baselines/ppo/ppo.py
index 8127307..bbf880b 100644
--- a/torchy_baselines/ppo/ppo.py
+++ b/torchy_baselines/ppo/ppo.py
@@ -253,6 +253,7 @@ class PPO(BaseRLModel):
         logger.logkv("entropy", entropy.mean().item())
         logger.logkv("policy_loss", policy_loss.item())
         logger.logkv("value_loss", value_loss.item())
+        logger.logkv("std", th.exp(self.policy.log_std).mean().item())
 
     def learn(self, total_timesteps, callback=None, log_interval=1,
               eval_env=None, eval_freq=-1, n_eval_episodes=5, tb_log_name="PPO", reset_num_timesteps=True):