diff --git a/jax/_src/pallas/mosaic/sc_core.py b/jax/_src/pallas/mosaic/sc_core.py
index 1f53daf96da3..0d95f24431ea 100644
--- a/jax/_src/pallas/mosaic/sc_core.py
+++ b/jax/_src/pallas/mosaic/sc_core.py
@@ -16,8 +16,9 @@
 from __future__ import annotations
 
 import collections
-from collections.abc import Sequence
+from collections.abc import Mapping, Sequence
 import dataclasses
+import functools
 import math
 from typing import Any, TypeAlias
 
@@ -340,8 +341,9 @@ def wrapper(*args):
       @pallas_core.core_map(mesh, **kwargs)
       def _():
         return pallas_primitives.run_scoped(
-            lambda *scratch_refs: body(*arg_refs, *out_refs, *scratch_refs),
-            *scratch_shapes,
+            functools.partial(body, *arg_refs, *out_refs),
+            *scratch_shapes if isinstance(scratch_shapes, Sequence) else (),
+            **scratch_shapes if isinstance(scratch_shapes, Mapping) else {},
         )
 
       outs = jax.tree.map(lambda ref: ref[...], out_refs)
diff --git a/tests/pallas/tpu_sparsecore_pallas_test.py b/tests/pallas/tpu_sparsecore_pallas_test.py
index 9b027a4a22da..491c51e561e8 100644
--- a/tests/pallas/tpu_sparsecore_pallas_test.py
+++ b/tests/pallas/tpu_sparsecore_pallas_test.py
@@ -1145,10 +1145,12 @@ def test_barrier_via_pallas_call(self):
             shape=(mesh.num_subcores, vec_dim), dtype=jnp.uint32
         ),
         out_specs=pl.BlockSpec((1, vec_dim), lambda i: (i, 0)),
-        scratch_shapes=[
-            pltpu.VMEM_SHARED((mesh.num_subcores, vec_dim), jnp.uint32),
-            pltpu.VMEM((vec_dim,), jnp.uint32),
-        ],
+        scratch_shapes=dict(
+            shared_ref=pltpu.VMEM_SHARED(
+                (mesh.num_subcores, vec_dim), jnp.uint32
+            ),
+            vmem_ref=pltpu.VMEM((vec_dim,), jnp.uint32),
+        ),
     )
     def kernel(o_ref, shared_ref, vmem_ref):
       subcore_id = pl.program_id(0)