diff --git a/config/training/dqn-hyperopt.toml b/config/training/dqn-hyperopt.toml
index e7cb50794..478f6ad86 100644
--- a/config/training/dqn-hyperopt.toml
+++ b/config/training/dqn-hyperopt.toml
@@ -139,7 +139,7 @@ exit_timing_weight = 0.05
 ofi_reward_weight = 0.2
 kelly_sizing_weight = 0.1
 reward_noise_scale = 0.05
-exposure_aux_weight = 0.01
+exposure_aux_weight = 0.1
 exposure_aux_warmup_epochs = 5
 
 [fixed]
@@ -182,5 +182,5 @@ exit_timing_weight = 0.05
 ofi_reward_weight = 0.2
 kelly_sizing_weight = 0.1
 reward_noise_scale = 0.05
-exposure_aux_weight = 0.01
+exposure_aux_weight = 0.1
 exposure_aux_warmup_epochs = 5
diff --git a/config/training/dqn-localdev.toml b/config/training/dqn-localdev.toml
index dacb6f82f..7ebd67609 100644
--- a/config/training/dqn-localdev.toml
+++ b/config/training/dqn-localdev.toml
@@ -131,5 +131,5 @@ exit_timing_weight = 0.05
 ofi_reward_weight = 0.2
 kelly_sizing_weight = 0.1
 reward_noise_scale = 0.05
-exposure_aux_weight = 0.01
+exposure_aux_weight = 0.1
 exposure_aux_warmup_epochs = 5
diff --git a/config/training/dqn-production.toml b/config/training/dqn-production.toml
index f10e07db7..de179d569 100644
--- a/config/training/dqn-production.toml
+++ b/config/training/dqn-production.toml
@@ -142,5 +142,5 @@ exit_timing_weight = 0.05
 ofi_reward_weight = 0.2
 kelly_sizing_weight = 0.1
 reward_noise_scale = 0.05
-exposure_aux_weight = 0.01
+exposure_aux_weight = 0.1
 exposure_aux_warmup_epochs = 5
diff --git a/config/training/dqn-smoketest.toml b/config/training/dqn-smoketest.toml
index 7ad05e31e..a36ea24b9 100644
--- a/config/training/dqn-smoketest.toml
+++ b/config/training/dqn-smoketest.toml
@@ -138,7 +138,7 @@ exit_timing_weight = 0.05
 ofi_reward_weight = 0.2
 kelly_sizing_weight = 0.1
 reward_noise_scale = 0.05
-exposure_aux_weight = 0.01
+exposure_aux_weight = 0.1
 exposure_aux_warmup_epochs = 5
 
 [walk_forward]
diff --git a/crates/ml/src/cuda_pipeline/gpu_dqn_trainer.rs b/crates/ml/src/cuda_pipeline/gpu_dqn_trainer.rs
index 67db881e8..e0c0ed457 100644
--- a/crates/ml/src/cuda_pipeline/gpu_dqn_trainer.rs
+++ b/crates/ml/src/cuda_pipeline/gpu_dqn_trainer.rs
@@ -2110,7 +2110,7 @@ impl GpuDqnTrainer {
 
         let goff_w_b0out: usize = sizes[..10].iter().sum::<usize>();
         let aux_params_ptr = self.ptrs.params_f32_ptr + (goff_w_b0out as u64) * 4;
-        let aux_lr = 0.001_f32;       // higher LR for aux (independent from main)
+        let aux_lr = 0.01_f32;        // 10x main LR — aux needs strong signal to break i%3
         let aux_beta1 = 0.9_f32;
         let aux_beta2 = 0.999_f32;
         let aux_eps = 1e-8_f32;
diff --git a/crates/ml/src/trainers/dqn/config.rs b/crates/ml/src/trainers/dqn/config.rs
index 247ed1851..eed0f87db 100644
--- a/crates/ml/src/trainers/dqn/config.rs
+++ b/crates/ml/src/trainers/dqn/config.rs
@@ -1653,7 +1653,7 @@ impl DQNHyperparameters {
             ofi_reward_weight: 0.2,
             kelly_sizing_weight: 0.1,
             reward_noise_scale: 0.05,
-            exposure_aux_weight: 0.01,
+            exposure_aux_weight: 0.1,
             exposure_aux_warmup_epochs: 5,
             // v8: Dense micro-reward, TD(λ), PopArt, curriculum, hindsight
             micro_reward_scale: 0.01,