diff --git a/alfredo/agents/aant/aant.py b/alfredo/agents/aant/aant.py
index 9003125..f65b796 100644
--- a/alfredo/agents/aant/aant.py
+++ b/alfredo/agents/aant/aant.py
@@ -157,12 +157,12 @@ def step(self, state: State, action: jax.Array) -> State:
         ctrl_cost = rControl_act_ss(self.sys,
                                     pipeline_state,
                                     action,
-                                    weight=0.0)
+                                    weight=-self._ctrl_cost_weight)
         
         torque_cost = rTorques(self.sys,
                                pipeline_state,
                                action,
-                               weight=0.0)        
+                               weight=-0.003)
 
         upright_reward = rUpright(self.sys,
                                   pipeline_state,
diff --git a/experiments/AAnt-locomotion/training.py b/experiments/AAnt-locomotion/training.py
index 182b0cc..fe40455 100644
--- a/experiments/AAnt-locomotion/training.py
+++ b/experiments/AAnt-locomotion/training.py
@@ -28,7 +28,7 @@
         "backend": "positional",
         "seed": 1,
         "len_training": 1_500_000,
-        "num_evals": 200,
+        "num_evals": 500,
         "num_envs": 2048,
         "batch_size": 2048,
         "num_minibatches": 8,
@@ -81,7 +81,10 @@ def progress(num_steps, metrics):
 
 scenes_fp = os.path.dirname(scenes.__file__)
 
-env_xml_paths = [f"{scenes_fp}/flatworld/flatworld_A1_env.xml"]
+env_xml_paths = [f"{scenes_fp}/flatworld/flatworld_A1_env.xml", 
+                 f"{scenes_fp}/flatworld/flatworld_A1_env.xml",
+                 f"{scenes_fp}/flatworld/flatworld_A1_env.xml",
+                 f"{scenes_fp}/flatworld/flatworld_A1_env.xml"]
 
 # make and save initial ppo_network
 key = jax.random.PRNGKey(wandb.config.seed)
@@ -115,7 +118,7 @@ def progress(num_steps, metrics):
 # ============================
 # Training & Saving Params
 # ============================
-i = 0
+i = 8
 
 for p in env_xml_paths:
 
diff --git a/experiments/AAnt-locomotion/vis_traj.py b/experiments/AAnt-locomotion/vis_traj.py
index 603c178..b1df220 100644
--- a/experiments/AAnt-locomotion/vis_traj.py
+++ b/experiments/AAnt-locomotion/vis_traj.py
@@ -79,7 +79,7 @@
 #yaw_vel = 0.0   # rad/s
 #jcmd = jp.array([x_vel, y_vel, yaw_vel])
 
-wcmd = jp.array([-10.0, 10.0])
+wcmd = jp.array([10.0, 10.0])
 
 # generate policy rollout
 for _ in range(episode_length):