FIx bug in the optimized cpu ops configuration (#2349)

Summary:
Pull Request resolved: #2349

Previous diff erroneously used portable for optimized dep
ghstack-source-id: 218361719
exported-using-ghexport

Reviewed By: JacobSzwejbka

Differential Revision: D54766069

fbshipit-source-id: 048c284c5a2047c80a4763c08ec5872576a78969

examples/models/llama2/runner/targets.bzl

@@ -6,7 +6,7 @@ def _get_operator_lib(aten = False):
     elif runtime.is_oss:
         return ["//executorch/kernels/portable:generated_lib"]
     else:
-        return ["//executorch/kernels/portable:generated_lib", "//executorch/examples/models/llama2/custom_ops:custom_ops", "//executorch/examples/models/llama2/ops:generated_lib"]
+        return ["//executorch/configurations:optimized_native_cpu_ops", "//executorch/examples/models/llama2/custom_ops:custom_ops", "//executorch/examples/models/llama2/ops:generated_lib"]
 
 def define_common_targets():
     for aten in (True, False):