diff --git a/developers/compiler/minituring-compiler/index.qmd b/developers/compiler/minituring-compiler/index.qmd
index c22894b17..0ac5ab9b5 100755
--- a/developers/compiler/minituring-compiler/index.qmd
+++ b/developers/compiler/minituring-compiler/index.qmd
@@ -270,7 +270,18 @@ We define the probabilistic model:
 end;
 ```
 
-We perform inference with data `x = 3.0`:
+The `@mini_model` macro expands this into another function, `m`, which effectively calls either `assume` or `observe` on each variable as needed:
+
+```{julia}
+@macroexpand @mini_model function m(x)
+    a ~ Normal(0.5, 1)
+    b ~ Normal(a, 2)
+    x ~ Normal(b, 0.5)
+    return nothing
+end
+```
+
+We can use this function to construct the `MiniModel`, and then perform inference with data `x = 3.0`:
 
 ```{julia}
 sample(MiniModel(m, (x=3.0,)), MHSampler(), 1_000_000; chain_type=Chains, progress=false)