diff --git a/jax/_src/pallas/mosaic/lowering.py b/jax/_src/pallas/mosaic/lowering.py
index 8d549d767da3..d6c30be8946f 100644
--- a/jax/_src/pallas/mosaic/lowering.py
+++ b/jax/_src/pallas/mosaic/lowering.py
@@ -2779,7 +2779,7 @@ def _rem_lowering_rule(ctx: LoweringRuleContext, x, y):
   raise NotImplementedError(aval_out.dtype)
 
 
-@register_lowering_rule(lax.abs_p)
+@register_lowering_rule(lax.abs_p, kernel_types=[*tpu_core.KernelType])
 def _abs_lowering_rule(ctx: LoweringRuleContext, x):
   (aval_out,) = ctx.avals_out
   if jnp.issubdtype(aval_out.dtype, jnp.integer):
@@ -2789,7 +2789,9 @@ def _abs_lowering_rule(ctx: LoweringRuleContext, x):
   raise NotImplementedError(aval_out.dtype)
 
 
-@register_lowering_rule(lax.neg_p, ensure_mlir_values=False)
+@register_lowering_rule(
+    lax.neg_p, kernel_types=[*tpu_core.KernelType], ensure_mlir_values=False
+)
 def _neg_lowering_rule(ctx: LoweringRuleContext, x):
   (x_aval,) = ctx.avals_in
   new_ctx = ctx.replace(
diff --git a/tests/pallas/tpu_sparsecore_pallas_test.py b/tests/pallas/tpu_sparsecore_pallas_test.py
index 2878e6ed1e6a..e43706f85899 100644
--- a/tests/pallas/tpu_sparsecore_pallas_test.py
+++ b/tests/pallas/tpu_sparsecore_pallas_test.py
@@ -1614,14 +1614,16 @@ def kernel(in_ref, o_ref, scratch_ref):
 
     np.testing.assert_array_equal(f(x), x)
 
-  def test_exp(self):
+  @parameterized.named_parameters(
+      ("exp", jnp.exp), ("neg", lambda x: -x), ("abs", jnp.abs))
+  def test_unary_ops(self, op):
     if not jtu.if_cloud_tpu_at_least(2025, 11, 30):
       self.skipTest("Test requires a newer libtpu")
 
     x = jnp.arange(8, dtype=jnp.float32)
 
     def sc_exp_kernel(x_hbm_ref, out_ref):
-      out_ref[...] = jnp.exp(x_hbm_ref[...])
+      out_ref[...] = op(x_hbm_ref[...])
 
     result = pl.pallas_call(
         sc_exp_kernel,
@@ -1630,7 +1632,7 @@ def sc_exp_kernel(x_hbm_ref, out_ref):
             ),
         out_shape=x,
     )(x)
-    np.testing.assert_array_equal(result, jnp.exp(x))
+    np.testing.assert_array_equal(result, op(x))
 
   @parameterized.product(dtype=[np.int32, np.float32])
   def test_vector_gather(self, dtype):