diff --git a/notebooks/jax_likelihood_functions/interferometer/rectangular_sparse.ipynb b/notebooks/jax_likelihood_functions/interferometer/rectangular_sparse.ipynb
index 91660f23..a4ccd793 100644
--- a/notebooks/jax_likelihood_functions/interferometer/rectangular_sparse.ipynb
+++ b/notebooks/jax_likelihood_functions/interferometer/rectangular_sparse.ipynb
@@ -333,7 +333,7 @@
         "\n",
         "np.testing.assert_allclose(\n",
         "    np.array(result),\n",
-        "    -3152.03184792,\n",
+        "    -3164.286252,\n",
         "    rtol=1e-4,\n",
         "    err_msg=\"interferometer/rectangular_sparse: JAX vmap likelihood mismatch\",\n",
         ")\n"
diff --git a/scripts/jax_assertions/sparse_operators.py b/scripts/jax_assertions/sparse_operators.py
index c30fb541..55b1271a 100644
--- a/scripts/jax_assertions/sparse_operators.py
+++ b/scripts/jax_assertions/sparse_operators.py
@@ -235,18 +235,29 @@
 
 pix_weights_for_sub_slim_index = np.ones(shape=(9, 1))
 
+# Use the unmasked-extent index for the interferometer sparse operator (which
+# lives on the extent grid, not the full native grid). For this fully-unmasked
+# 3x3 grid the two indices are identical, but the extent form is the convention
+# accepted by the operator after the Pmax > 1 fix.
+rows_interferometer, cols_interferometer, vals_interferometer = aa.util.mapper.sparse_triplets_from(
+    pix_indexes_for_sub=pix_indexes_for_sub_slim_index,
+    pix_weights_for_sub=pix_weights_for_sub_slim_index,
+    slim_index_for_sub=np.arange(pix_indexes_for_sub_slim_index.shape[0], dtype=np.int32),
+    fft_index_for_masked_pixel=grid.mask.extent_index_for_masked_pixel,
+    sub_fraction_slim=np.ones(pix_indexes_for_sub_slim_index.shape[0], dtype=np.float64),
+    return_rows_slim=False,
+)
+
 sparse_operator = aa.InterferometerSparseOperator.from_nufft_precision_operator(
     nufft_precision_operator=nufft_precision_operator,
     dirty_image=None,
 )
 
-curvature_matrix_via_preload = (
-    sparse_operator.curvature_matrix_via_sparse_operator_from(
-        pix_indexes_for_sub_slim_index=pix_indexes_for_sub_slim_index,
-        pix_weights_for_sub_slim_index=pix_weights_for_sub_slim_index,
-        fft_index_for_masked_pixel=grid.mask.fft_index_for_masked_pixel,
-        pix_pixels=3,
-    )
+curvature_matrix_via_preload = sparse_operator.curvature_matrix_diag_from(
+    rows=rows_interferometer,
+    cols=cols_interferometer,
+    vals=vals_interferometer,
+    S=3,
 )
 
 npt.assert_allclose(
diff --git a/scripts/jax_likelihood_functions/interferometer/rectangular_sparse.py b/scripts/jax_likelihood_functions/interferometer/rectangular_sparse.py
index a670a4a4..57f6525f 100644
--- a/scripts/jax_likelihood_functions/interferometer/rectangular_sparse.py
+++ b/scripts/jax_likelihood_functions/interferometer/rectangular_sparse.py
@@ -213,7 +213,7 @@ class in **PyAutoFit**, which pairs the model with likelihood.
 
 np.testing.assert_allclose(
     np.array(result),
-    -3152.03184792,
+    -3164.286252,
     rtol=1e-4,
     err_msg="interferometer/rectangular_sparse: JAX vmap likelihood mismatch",
 )
@@ -263,12 +263,12 @@ class in **PyAutoFit**, which pairs the model with likelihood.
 The Path A run above uses TransformerDFT + `apply_sparse_operator(use_jax=True)`
 (the cached-precision-matrix accelerator for pixelization). This pass uses
 the same TransformerDFT but skips the sparse-operator optimization — the
-plain direct-DFT pixelization path. The two paths give *different*
-log-likelihoods at the ~0.4% level because the sparse-operator
-precomputation is a numerical reformulation, not an exact reproduction;
-this is expected. Path B's literal must therefore match
-`scripts/jax_likelihood_functions/interferometer/rectangular.py` (which is
-the same model + DFT-no-sparse path).
+plain direct-DFT pixelization path. After the Pmax > 1 / extent-indexing fix
+(issue #314), the two paths agree to numerical precision: the sparse-operator
+precomputation is mathematically exact, not a "numerical reformulation".
+Path B's literal therefore matches Path A and
+`scripts/jax_likelihood_functions/interferometer/rectangular.py` (the same
+model + DFT-no-sparse path).
 """
 dataset_dft_nosparse = al.Interferometer.from_fits(
     data_path=path.join(dataset_path, "data.fits"),