diff --git a/test/chainrules.jl b/test/chainrules.jl
index 0ab7cdb..6023ac3 100644
--- a/test/chainrules.jl
+++ b/test/chainrules.jl
@@ -7,39 +7,38 @@ using ChainRulesTestUtils
     y = randn(n)
     X = randn(n, 3)
     Y = randn(n, 3)
+    Xrep = repeat(x, 1, 3)
+    Yrep = repeat(y, 1, 3)
 
     @testset for metric in (SqEuclidean(), Euclidean())
-        @testset "different arguments" begin
-            # Single evaluation
-            test_rrule(metric ⊢ NoTangent(), x, y)
+        # Single evaluation
+        test_rrule(metric ⊢ NoTangent(), x, y)
+        test_rrule(metric ⊢ NoTangent(), x, x)
 
+        for A in (X, Xrep)
             # Column-wise distance
-            test_rrule(colwise, metric ⊢ NoTangent(), X, Y)
-
-            # Pairwise distances
-            test_rrule(pairwise, metric ⊢ NoTangent(), X)
-            test_rrule(pairwise, metric ⊢ NoTangent(), X; fkwargs=(dims=1,))
-            test_rrule(pairwise, metric ⊢ NoTangent(), X; fkwargs=(dims=2,))
-            test_rrule(pairwise, metric ⊢ NoTangent(), X, Y)
-            test_rrule(pairwise, metric ⊢ NoTangent(), X, Y; fkwargs=(dims=1,))
-            test_rrule(pairwise, metric ⊢ NoTangent(), X, Y; fkwargs=(dims=2,))    
-        end
-
-        # check numerical issues if distances are zero
-        @testset "equal arguments" begin
-            # Single evaluation
-            test_rrule(metric ⊢ NoTangent(), x, x)
-
-            # Column-wise distance
-            test_rrule(colwise, metric ⊢ NoTangent(), X, X)
+            test_rrule(colwise, metric ⊢ NoTangent(), A, A)
 
             # Pairwise distances
             # Finite differencing yields impressively inaccurate derivatives for `Euclidean`,
             # see https://github.com/FluxML/Zygote.jl/blob/45bf883491d2b52580d716d577e2fa8577a07230/test/gradcheck.jl#L1206
-            kwargs = metric isa Euclidean ? (rtol = 1e-4,) : ()
-            test_rrule(pairwise, metric ⊢ NoTangent(), X, X; kwargs...)
-            test_rrule(pairwise, metric ⊢ NoTangent(), X, X; fkwargs=(dims=1,), kwargs...)
-            test_rrule(pairwise, metric ⊢ NoTangent(), X, X; fkwargs=(dims=2,), kwargs...)
+            kwargs = metric isa Euclidean ? (rtol=1e-3, atol=1e-3) : ()
+            test_rrule(pairwise, metric ⊢ NoTangent(), A; kwargs...)
+            test_rrule(pairwise, metric ⊢ NoTangent(), A; fkwargs=(dims=1,), kwargs...)
+            test_rrule(pairwise, metric ⊢ NoTangent(), A; fkwargs=(dims=2,), kwargs...)
+            test_rrule(pairwise, metric ⊢ NoTangent(), A, A; kwargs...)
+            test_rrule(pairwise, metric ⊢ NoTangent(), A, A; fkwargs=(dims=1,), kwargs...)
+            test_rrule(pairwise, metric ⊢ NoTangent(), A, A; fkwargs=(dims=2,), kwargs...)
+
+            for B in (Y, Yrep)
+                # Column-wise distance
+                test_rrule(colwise, metric ⊢ NoTangent(), A, B)
+
+                # Pairwise distances
+                test_rrule(pairwise, metric ⊢ NoTangent(), A, B)
+                test_rrule(pairwise, metric ⊢ NoTangent(), A, B; fkwargs=(dims=1,))
+                test_rrule(pairwise, metric ⊢ NoTangent(), A, B; fkwargs=(dims=2,))
+            end
         end
     end
 end