diff --git a/infra/k8s/argo/alpha-rl-template.yaml b/infra/k8s/argo/alpha-rl-template.yaml
index 5a00bfcc4..e32a328c9 100644
--- a/infra/k8s/argo/alpha-rl-template.yaml
+++ b/infra/k8s/argo/alpha-rl-template.yaml
@@ -61,6 +61,8 @@ spec:
         value: "0"
       - name: nsys-profile
         value: "false"
+      - name: use-multi-head-policy
+        value: "0"
 
   volumes:
     - name: git-ssh-key
@@ -102,6 +104,8 @@ spec:
             value: /cargo-target/cargo-home
           - name: CARGO_INCREMENTAL
             value: "1"
+          - name: FOXHUNT_USE_MULTI_HEAD_POLICY
+            value: "{{workflow.parameters.use-multi-head-policy}}"
         resources:
           requests:
             cpu: "4"
diff --git a/scripts/argo-alpha-rl.sh b/scripts/argo-alpha-rl.sh
index b0b9dc0f3..6fe67c3f0 100755
--- a/scripts/argo-alpha-rl.sh
+++ b/scripts/argo-alpha-rl.sh
@@ -46,6 +46,7 @@ N_EVAL_STEPS=0
 WATCH=false
 SKIP_PUSH_CHECK=false
 SKIP_TEMPLATE_APPLY=false
+USE_MULTI_HEAD_POLICY=0
 
 usage() {
   cat <<EOF
@@ -109,6 +110,7 @@ while [[ $# -gt 0 ]]; do
     --watch)               WATCH=true; shift ;;
     --skip-push-check)     SKIP_PUSH_CHECK=true; shift ;;
     --skip-template-apply) SKIP_TEMPLATE_APPLY=true; shift ;;
+    --use-multi-head-policy) USE_MULTI_HEAD_POLICY="$2"; shift 2 ;;
     -h|--help)             usage; exit 0 ;;
     *) echo "Unknown option: $1"; usage; exit 1 ;;
   esac
@@ -221,4 +223,5 @@ argo submit -n foxhunt --from=wftmpl/alpha-rl \
   -p fold-idx="$FOLD_IDX" \
   -p n-folds="$N_FOLDS" \
   -p n-eval-steps="$N_EVAL_STEPS" \
+  -p use-multi-head-policy="$USE_MULTI_HEAD_POLICY" \
   $WATCH_FLAG