diff --git a/modules/sampling/noise_merge/__init__.py b/modules/sampling/noise_merge/__init__.py
index d8e94a48..ef1fbba2 100644
--- a/modules/sampling/noise_merge/__init__.py
+++ b/modules/sampling/noise_merge/__init__.py
@@ -101,11 +101,19 @@ def install(patcher, values, key):
         return None
     b_first = values.get("order", "b_into_a") == "b_into_a"
     chosen = int(values.get("seed", -1))
-    live = {"seed": max(chosen, 0)}
+    live = {"seed": max(chosen, 0), "merged": 0}
 
     def start(executor, *args, **kwargs):
         live["seed"] = chosen if chosen >= 0 else random.randrange(2 ** 31)
-        return executor(*args, **kwargs)
+        live["merged"] = 0
+        log.info("FunUI Noise merging", f"on: steps {values.get('steps')}, strength {strength:.2f}, "
+                 f"second seed {live['seed']}")
+        try:
+            return executor(*args, **kwargs)
+        finally:
+            if not live["merged"]:
+                _say("Inactive | no step of this run matched the steps you named "
+                     "(or the model was not called as a plain prompt step)")
 
     patcher.add_wrapper_with_key(WrappersMP.OUTER_SAMPLE, key, start)
 
@@ -132,6 +140,7 @@ def install(patcher, values, key):
             if split_b is None:
                 return out
             video_b, rebuild_b = split_b
+            live["merged"] += 1
             diff = float((video_a - video_b).norm() / video_a.norm().clamp(min=1e-8))
             log.info("FunUI Noise merging",
                      f"step {where[0] + 1}: second answer differs from the first by {diff:.1%}, "
diff --git a/modules/sampling/noise_merge/tests/test_noise_merge.py b/modules/sampling/noise_merge/tests/test_noise_merge.py
index f45004ff..ad2ad68a 100644
--- a/modules/sampling/noise_merge/tests/test_noise_merge.py
+++ b/modules/sampling/noise_merge/tests/test_noise_merge.py
@@ -106,3 +106,19 @@ def test_a_negative_seed_is_new_each_run_and_a_number_is_kept(tiny_h3):
         for _ in range(2):
             outer(lambda: wrap(model, x0.clone(), _sched()[0:1], None, None, None, _to(0), latent_shapes=shapes))
         assert torch.equal(seen[1], seen[3]) is same
+
+
+def test_the_log_says_the_feature_is_on_and_when_nothing_matched(tiny_h3):
+    from comfy.patcher_extension import WrappersMP
+    from core import log
+    from modules.sampling.modifiers.nodes import FunUILoadModifiers
+    x0, shapes = _packed()
+    patched, _ = FunUILoadModifiers.execute(
+        tiny_h3.patcher, {"noise_merge": {"enabled": True, "steps": "9"}}).result
+    wrap = [w for ws in patched.wrappers[WrappersMP.APPLY_MODEL].values() for w in ws][-1]
+    outer = [w for ws in patched.wrappers[WrappersMP.OUTER_SAMPLE].values() for w in ws][-1]
+    log.new_run()
+    outer(lambda: wrap(lambda *a, **k: x0, x0.clone(), _sched()[0:1], None, None, None, _to(0), latent_shapes=shapes))
+    texts = [e["message"] for e in log.history()]
+    assert any(t.startswith("on: steps 9") for t in texts)
+    assert any("no step of this run matched" in t for t in texts)