diff --git a/src/interface.jl b/src/interface.jl
index 9b199e0a..2c0ae7c3 100644
--- a/src/interface.jl
+++ b/src/interface.jl
@@ -127,7 +127,7 @@ end
 # N.B: This means the gradient is not available for zero-arg function, but such
 # a gradient would be guaranteed to be `()`, which is a bit of a useless thing
 function (::Type{∇})(f, x1, args...)
-    ∇(f)(x1, args...)
+    unthunk.(∇(f)(x1, args...))
 end
 
 const gradient = ∇
@@ -159,7 +159,7 @@ function (f::PrimeDerivativeBack)(x)
     z = ∂⃖¹(lower_pd(f), x)
     y = getfield(z, 1)
     f☆ = getfield(z, 2)
-    return getfield(f☆(dx(y)), 2)
+    return unthunk(getfield(f☆(dx(y)), 2))
 end
 
 # Forwards primal derivative
diff --git a/test/runtests.jl b/test/runtests.jl
index c7ae67d8..4b1832e6 100644
--- a/test/runtests.jl
+++ b/test/runtests.jl
@@ -108,7 +108,7 @@ let var"'" = Diffractor.PrimeDerivativeBack
     # Control flow cases
     @test @inferred((x->simple_control_flow(true, x))'(1.0)) == sin'(1.0)
     @test @inferred((x->simple_control_flow(false, x))'(1.0)) == cos'(1.0)
-    @test_broken (x->sum(isa_control_flow(Matrix{Float64}, x)))'(Float32[1 2;]) == [1.0 1.0;]
+    @test (x->sum(isa_control_flow(Matrix{Float64}, x)))'(Float32[1 2;]) == [1.0 1.0;]
     @test times_three_while'(1.0) == 3.0
 
     pow5p(x) = (x->mypow(x, 5))'(x)